Drago - Physicalmethodsforchemists 2e PDF

Physical
Methods for
Chemists
SECO ND ED ITIO N
RUSSELL S. PRAGO
University of Florida, Gainesville
9,RFSIDE
SCIENTIFIC PUBLISHERS
P.O. Box 13413

Gainesville, FL 32604
"7
!
Copyright @ 1992, 1977

All rights reserved. No part of this publication may be reproduced or transmitted
in any form or by any means, electronic or mechanical, including photocopy,
recording or any information storage and retrieval system, without permission in
writing from the publisher.
Requests for permission to make copies of any part of the work should be mailed
to: Surfside Scientific Publishers, P.O. Box 13413, Gainesville, FL 32604-1413.
Text Typeface: Times Roman
Library of Congress Cataloging-in-Publication Data
Drago, Russell S.
Physical methods for chemists / Russell S. Drago. - 2nd ed. p. cm.
Rev. ed. of: Physical methods in chemistry. 1977.
Includes bibliographical references (p. ) and index.
ISBN 0-03-075176-4
1. Chemistry, Physical and theoretical. 2. Spectrum analysis.
I. Drago, Russell S. Physical methods in chemistry. II. Title.
QD453.2.D7 1992
543--dc2O 91-41550
CIP
OTHER SURFSIDE TITLES:
SOLUTION MANUAL - PHYSICAL METHODS FOR CHEMISTS by Ngai M.

Wong and Russell S. Drago
APPLICATIONS OF ELECTROSTATIC-COVALENT MODELS IN

CHEMISTRY by Russell S. Drago
M..T. LIBRARIES Send orders to Surfside Scientific Publishers, P.O. Box 13413, Gainesville, FL
32604-1413
DE'C 14 2005
RECEIVED
To RUTH
with love
Preface
The revision of my earlier textbook, "Physical Methods in Chemistry," was

motivated by a sense of gratitude to those who continued to use the book even
though several parts were out of date. Those parts of the last edition that remain
current have not been changed. As in the earlier edition, topics have been
subdivided into numbered paragraphs so some can be dropped, if desired,
permitting use of the text in either a theoretical or more applications-oriented
course.
The popularity of the earlier edition is based on the recognition that there
is more to the application of physical methods in chemistry than memorizing
where functional groups appear. Core material in the area of spectroscopy is
presented that is of importance to both organic and inorganic chemists. This
text is designed to reflect this unity. A fundamental subdivision of chemistry
involves dividing it into systems that do and do not contain unpaired d-electrons.
The material in this book has been so subdivided. In using this book, a chemist
with an interest in organic or organometallic sysnthesis can omit Chapters 10,
11, 12, and 13.
This book contains more material than can be covered reasonably in a
one-semester course. Accordingly, several methods in less common usage (Moss-
bauer, NQR, surface techniques) are presented in a succinct, basic manner. In
this way, the reader is at least made aware of their existence.
The book continues in the philosophy of the earlier edition, namely, that
chemists without an advanced mathematical background learn to use spectro-
scopic methods by reading about how problems have been solved with them.
Thus, the text relies heavily on examples which illustrate the kinds of information
that can be obtained by application of the methods. Whenever a choice exists
regarding the selection of an application, an example from my own research has
been chosen, for I am most familiar with the details of this work. I recommend
that instructors supplement this material with examples reflecting their particular
interests. Hopefully, more widespread knowledge of the core material in this text
will upgrade or at least make the reader more critical of conclusions from
spectroscopic measurements. I feel this is one of the important contributions of
the earlier text and the original motivation for writing this book.
This text has evolved from the author's presentation of courses in physical
methods. He gratefully acknowledges the contributions to the development
of this material by the faculty, teaching assistants, and students involved,
VI Preface
particularly those mentioned in earlier editions. I appreciate very much the fine
contributions to this edition by Ivano Bertini, Claudio Luchinat, Arnie Rhein-
gold, and Joe Ziller, who found the time to help a friend. Thanks go to Kitty
Williams for Figure 8-25, Dave Powell for suggestions on the mass spectroscopy
section, and to Roy King for many valuable comments and a very thorough
review of Chapters 7 and 8. Thanks go to Steve Showalter, Steve Petrosius, and
Mike Robbins for hours of proofreading. This revision would not have occurred
if it were not for the help of "super secretary" Maribel Lisk and her able assistant
April Kirch. I appreciate their dedication, the pride they take in a job well done,
and their tolerance of a hyper, "I want it yesterday," professor. I also thank
Dyana Drago for help in preparing the manuscript, but most of all, for
collaborating with son Steve and giving us (Ruth and I) granddaughters Stacie
and Olivia. John Vondeling's faith in the value of this effort is greatly appreciated,
with the fond hope that, for this edition, he will provide a cover that does not
self-destruct in a year. Finally, I would like to apologize to my wife, Ruth, for
the many hours of my time spent on chemistry instead of with her and thank
her for encouragement and unselfishness.
RUSSELL S. DRAGO
Contents
1
Symmetry and the Point Groups 1
Introduction 1
1-1 Definition of Symmetry 1
1-2 Symmetry Elements 2
1-3 Point Groups 8
1-4 Space Symmetry 11
1-5 Some Definitions and Applications of Symmetry
Considerations 12
References Cited 14
Additional Reading 14
Exercises 14
2
Group Theory and the Character Tables 18
2-1 Introduction 18
2-2 Rules for Elements that Constitute a Group 19
2-3 Group Multiplication Tables 19
2-4 Summary of the Properties of Vectors and Matrices 25
2-5 Representations; Geometric Transformations 30
2-6 Irreducible Representations 34
2-7 Character Tables 35
2-8 Non-Diagonal Representations 36
2-9 More on Character Tables 41
2-10 More on Representations 43
2-11 Simplified Procedures for Generating and Factoring Total
Representations: The Decomposition Formula 44
2-12 Direct Products 46
Exercises 47
Viii Contents
3
Molecular Orbital Theory and Its Symmetry Aspects 52
Introduction 52
3-1 Operators 52
3-2 A Matrix Formulation of Molecular Orbital Calculations 56
3-3 Perturbation Theory 57
Symmetry in Quantum Mechanics 59
3-4 Wave Functions as a Basis for Irreducible Representations 59
3-5 Projecting Molecular Orbitals 60
Molecular Orbital Calculations 68
3-6 Hiickel Procedure 68
3-7 Properties Derived from Wave Functions 72
3-8 Extended Huckel Procedure 74
3-9 SCF-INDO (Intermediate Neglect of Differential Overlap) 78
3-10 Some Predictions from M.O. Theory on Alternately Double
Bonded Hydrocarbons 83
3-11 More on Product Ground State Wave Functions 84
References Cited 85
Compilations 86
Exercises 86
4
General Introduction to Spectroscopy 90
4-1 Nature of Radiation 90
4-2 Energies Corresponding to Various Kinds of Radiation 91
4-3 Atomic and Molecular Transitions 92
4-4 Selection Rules 95
4-5 Relaxation and Chemical Exchange Influences on Spectral
Line Width 95
General Applications 98
4-6 Determination of Concentration 99
4-7 Isosbestic Points 103
4-8 Job's Method of Isomolar Solutions 106
4-9 "Fingerprinting" 106
References Cited 107
Exercises 107
5
Electronic Absorption Spectroscopy 109
Introduction 109
5-1 Vibrational and Electronic Energy Levels in a Diatomic
Molecule 109
Contents iX
5-2 Relationship of Potential Energy Curves to Electronic Spectra 111

5-3 Nomenclature 113
Assignment of Transitions 117
5-4 Spin-Orbit Coupling 117
5-5 Configuration Interaction 117
5-6 Criteria to Aid in Band Assignment 118
The Intensity of Electronic Transitions 120
5-7 Oscillator Strengths 120
5-8 Transition Moment Integral 120
5-9 Derivation of Some Selection Rules 123
5-10 Spectrum of Formaldehyde 123
5-11 Spin-Orbit and Vibronic Coupling Contributions to Intensity 124
5-12 Mixing of d and p Orbitals in Certain Symmetries 126
5-13 Magnetic Dipole and Electric Quadrupole Contributions to
Intensity 127
5-14 Charge Transfer Transitions 127
5-15 Polarized Absorption Spectra 128
Applications 130
5-16 Fingerprinting 130
5-17 Molecular Addition Compounds of Iodine 133
5-18 Effect of Solvent Polarity on Charge-Transfer Spectra 135
5-19 Structures of Excited States 137
Optical Rotary Dispersion, Circular Dichroism, and
Magnetocircular Dichroism 137
5-20 Introduction 137
5-22 Applications 140
5-23 Magnetocircular Dichroism 141
Additional References 144
Exercises 145
6
Vibration and Rotation Spectroscopy: Infrared, Raman, and Microwave 149
Introduction 149
6-1 Harmonic and Anharmonic Vibrations 149
6-2 Absorption of Radiation by Molecular Vibrations-
Selection Rules 150
6-3 Force Constant 151
Vibrations in a Polyatomic Molecule 153
6-4 The 3N - 6(5) Rule 153
6-5 Effects Giving Rise to Absorption Bands 154
X Contents
6-6 Normal Coordinate Analyses and Band Assignments 156

6-7 Group Vibrations and the Limitations of This Idea 160
Raman Spectroscopy 162
6-9 Raman Selection Rules 164
6-10 Polarized and Depolarized Raman Lines 168
6-11 Resonance Raman Spectroscopy 170
Symmetry Aspects of Molecular Vibrations 172
6-12 Significance of the Nomenclature Used to Describe Various
Vibrations 172
6-13 Use of Symmetry Considerations to Determine the Number
of Active Infrared and Raman Lines 172
6-14 Symmetry Requirements for Coupling, Combination Bands,
and Fermi Resonance 176
6-15 Microwave Spectroscopy 177
6-16 Rotational Raman Spectra 179
Applications of Infrared and Raman Spectroscopy 179
6-17 Procedures 179
6-19 Spectra of Gases 186
6-20 Application of Raman and Infrared Selection Rules to the
Determination of Inorganic Structures 192
Bond Strength-Frequency Shift Relations 194
6-21 Changes in the Spectra of Donor Molecules upon
Coordination 196
6-22 Change in Spectra Accompanying Change in Symmetry
upon Coordination 198
Exercises 206
Nuclear Magnetic Resonance Spectroscopy-Elementary Aspects 211

Introduction 211
Classical Description of the NMR Experiment-The Bloch
Equations 212
7-1 Some Definitions 212
7-2 Behavior of a Bar Magnet in a Magnetic Field 213
7-3 Rotating Axis Systems 214
7-4 Magnetization Vectors and Relaxation 215
7-5 The NMR Transition 217
7-6 The Bloch Equations 218
7-7 The NMR Experiment 220
Contents XI
The Quantum Mechanical Description of the NMR Experiment 224

7-8 Properties of I 224
7-9 Transition Probabilities 225
Relaxation Effects and Mechanisms 227
7-10 Measuring the Chemical Shift 229
7-11 Interpretation of the Chemical Shift 232
7-12 Interatomic Ring Currents 241
7-13 Examples of Chemical Shift Interpretation 241
Spin-Spin Splitting 243
7-14 Effect of Spin-Spin Splitting on the Spectrum 243
7-15 Discovering Non-Equivalent Protons 246
7-16 Effect of the Number and Nature of the Bonds on
Spin-Spin Coupling 247
7-17 Scalar Spin-Spin Coupling Mechanisms 249
7-18 Application of Spin-Spin Coupling to Structure
Determination 252
Factors Influencing the Appearance of the NMR Spectrum 257
7-19 Effect of Fast Chemical Reactions on the Spectrum 257
7-20 Quantum Mechanical Description of Coupling 259
7-21 Effects of the Relative Magnitudes of J and A on the Spectrum
of an AB Molecule 263
7-22 More Complicated Second-Order Systems 265
7-23 Double Resonance and Spin-Tickling Experiments 267
7-24 Determining Signs of Coupling Constants 268
7-25 Effects on the Spectrum of Nuclei with Quadrupole Moments 269
Compilations of Chemical Shifts 272
Exercises 273
8
Dynamic and Fourier Transform NMR 290
Introduction 290
Evaluation of Theormodynamic Data with NMR 290
NMR Kinetics 291
8-1 Rate Constants and Activation Enthalpies from NMR 291
8-2 Determination of Reaction Orders by NMR 295
8-3 Some Applications of NMR Kinetic Studies 297
8-4 Intramolecular Rearrangements Studied by NMR-
Fluxional Behavior 300
8-5 Spin Saturation Labeling 305
8-6 The Nuclear Overhauser Effect 306
Xii Contents
Fourier Transform NMR 309

8-7 Principles 309
8-8 Optimizing the FTNMR Experiment 314
8-9 The Measurement of T by FTNMR 315
8-10 Use of T, for Peak Assignments 317
8-11 NMR of Quadrupolar Nuclei 319
Applications and Strategies in FTNMR 319
3 319
8-12 C
8-13 Other Nuclei 323
More on Relaxation Processes 326
8-14 Spectral Density 326
Multipulse Methods 328
8-16 Spin Echoes 329
8-17 Sensitivity-Enhancement Methods 331
8-18 Selective Excitation and Suppression 332
8-19 Two-Dimensional NMR 334
NMR in Solids and Liquid Crystals 340
8-20 Direct Dipolar Coupling 340
8-21 NMR Studies of Solids 341
8-22 NMR Studies in Liquid Crystal Solvents 342
8-23 High Resolution NMR of Solids 347
Exercises 352
9
Electron Paramagnetic Resonance Spectroscopy 360
Introduction 360
9-1 Principles 360
Nuclear Hyperfine Splitting 363
9-2 The Hydrogen Atom 363
9-3 Presentation of the Spectrum 368
9-4 Hyperfine Splittings in Isotropic Systems Involving More
Than One Nucleus 370
9-5 Contributions to the Hyperfine Coupling Constant in
Isotropic Systems 374
Anisotropic Effects 380
9-6 Anisotropy in the g Value 380
9-7 Anisotropy in the Hyperfine Coupling 383
9-8 The EPR of Triplet States 390
Contents Xii
9-9 Nuclear Quadrupole Interaction 392

9-10 Line Widths in EPR 394
9-11 The Spin Hamiltonian 396
9-12 Miscellaneous Applications 397
Exercises 401
10
The Electronic Structure and Spectra of Transition Metal Ions 409
Introduction 409
Free Ion Electronic States 409
10-1 Electron-Electron Interactions and Term Symbols 409
10-2 Spin-Orbit Coupling in Free Ions 413
Crystal Fields 416
10-3 Effects of Ligands on the d Orbital Energies 416
10-4 Symmetry Aspects of the d Orbital Splitting by Ligands 419
10-5 Double Groups 427
10-6 The Jahn-Teller Effect 429
10-7 Magnetic Coupling in Metal Ion Clusters 430
Applications 433
10-8 Survey of the Electronic Spectra of 0 Complexes 433
10-9 Calculation of Dq and # for 0 h Ni(II) Complexes 437
10-10 Effect of Distortions on the d Orbital Energy Levels 443
10-11 Structural Evidence from the Electronic Spectrum 447
Bonding Parameters from Spectra 450
10-12 a and 7E Bonding Parameters from the Spectra of
Tetragonal Complexes 450
10-13 The Angular Overlap Model 452
Miscellaneous Topics Involving Electronic Ttransitions 458
10-14 Electronic Spectra of Oxo-Bridged Dinuclear Iron Centers 458
10-15 Intervalence Electron Transfer Bands 459
10-16 Photoreactions 462
Exercises 463
11
Magnetism 469
11-2 Types of Magnetic Behavior 471
XiV Contents
11-3 Van Vleck's Equation 476

11-4 Applications of Susceptibility Measurements 483
11-5 Intramolecular Effects 486
11-6 High Spin-Low Spin Equilibria 489
11-7 Measurement of Magnetic Susceptibilities 490
11-8 Superparamagnetism 491
Exercises 496
12
Nuclear Magnetic Resonance of Paramagnetic Substances in Solution 500
by Ivano Bertini and C. Luchinat
12-2 Properties of Paramagnetic Compounds 501
12-3 Considerations Concerning Electron Spin 503
12-4 The Contact Shift 507
12-5 The Pseudocontact Shift 508
12-6 Lanthanides 510
12-7 Factoring the Contact and Pseudocontact Shifts 512
12-8 The Contact Shift and Spin Density 514
12-9 Factors Affecting Nuclear Relaxation in Paramagnetic
System 519
12-10 Relaxometry 524
12-11 Electronic Relaxation Times 527
12-12 Contrast Agents 530
12-13 Trends in the Development of Paramagnetic NMR 531
12-14 Some Applications 534
12-15 The Investigation of Bimetallic Systems 543
Exercises 556
13
Electron Paramagnetic Resonance Spectra of Transition Metal Ion
503
Complexes 559
13-2 Interpretation of the g-Values 564
13-3 Hyperfine Couplings and Zero Field Splittings 571
13-4 Ligand Hyperfine Couplings 576
13-5 Survey of the EPR Spectra of First-Row Transition Metal Ion
Complexes 578
13-6 The EPR of Metal Clusters 591
13-7 Double Resonance and Fourier Transform EPR Techniques 594
Exercises 597
Contents XV
14
Nuclear Quadrupole Resonance Spectroscopy, NOR 604
14-2 Energies of the Quadrupole Transitions 607
14-3 Effect of a Magnetic Field on the Spectra 611
14-4 Relationship Between Electric Field Gradient and
Molecular Structure 612
14-6 Double Resonance Techniques 620
Exercises 624
15
M6ssbauer Spectroscopy 626
15-2 Interpretation of Isomer Shifts 630
15-3 Quadrupole Interactions 631
15-4 Paramagnetic Mbssbauer Spectra 633
15-5 Mossbauer Emission Spectroscopy 635
Series 646
Exercises 647
16
Ionization Methods: Mass Spectrometry, Ion Cyclotron Resonance,
Photoelectron Spectroscopy 650
Mass Spectrometry 650
16-1 Instrument Operation and Presentation of Spectra 650
16-2 Processes That Can Occur When a Molecule and a High
Energy Electron Combine 655
16-3 Fingerprint Application 657
16-4 Interpretation of Mass Spectra 659
16-5 Effect of Isotopes on the Appearance of a Mass
Spectrum 660
16-6 Molecular Weight Determinations; Field Ionization
Techniques 662
16-7 Evaluation of Heats of Sublimation and Species in the
Vapor Over High Melting Solids 663
16-8 Appearance Potentials and Ionization Potentials 664
FTICR/MS 665
16-9 The Fourier Transform Ion Cyclotron Resonance Technique 665
XVi Contents
Surface Science Techniques 667

16-11 Photoelectron Spectroscopy 667
16-12 SIMS (Secondary Ion Mass Spectrometry) 677
16-13 LEED, AES, and HREELS Spectroscopy 678
16-14 STM (Scanning Tunneling Microscopy) and AFM
(Atomic Force Microscopy) 680
EXAFS and XANES 681
Exercises 685
17
X-Ray Crystallography 689
by Joseph W. Ziller and Arnold L. Rheingold

Principles 690
17-2 Diffraction of X-Rays 690
17-3 Reflection and Reciprocal Space 691
17-4 The Diffraction Pattern 693
17-5 X-Ray Scattering by Atoms and Structures 693
Crystals 695
17-6 Crystal Growth 695
17-7 Selection of Crystals 696
17-8 Mounting Crystals 697
Methodology 698
17-9 Diffraction Equipment 698
17-10 Diffractometer Data Collection 699
17-11 Computers 700
Some Future Developmenti- 700
17-12 Area Detectors 700
17-13 X-Ray versus Neutron Diffraction 701
17-14 Synchrotron Radiation 701
Symmetry and Related Concerns 701
17-15 Crystal Classes 701
17-16 Space Groups 703
17-17 Space-Group Determination 704
17-18 Avoiding Crystallographic Mistakes 705
17-19 Molecular versus Crystallographic Symmetry 707
17-20 Quality Assessment 707
17-21 Crystallographic Data 710
Contents XVIl

Exercises 711
Appendix A
Character Tables for Chemically Important Symmetry Groups 713
Appendix B
Character Tables for Some Double Groups 724
Appendix C
Normal Vibration Modes for Common Structures 727
Appendix D
0
Tanabe and Sugano Diagrams for h Fields 732
Appendix E
Calculation of A (1ODq) and # for Oh Ni" and Td Co" Complexes 735
Appendix F
Conversion of Chemical Shift Data 739
Appendix G
Solution of the Secular Determinant for the NMR Coupling of the AB Spin
System 740
Index
Symmetry and the
Point Groups
1-1 DEFINITION OF SYMMETRY Introduction

Symmetry considerations are fundamental to many areas of chemical reactivity,
electronic structure, and spectroscopy. It is customary to describe the structures
of molecules in terms of the symmetry that the molecules possess. Spectroscopists
have described molecular vibrations in terms of symmetry for many years. Modern
applications of spectroscopic methods to the problem of structure determination
require a knowledge of symmetry properties. We shall be concerned mainly with
the description of the symmetry of an isolated molecule, the so-called point
symmetry. Point symmetry refers to the set of operations transforming a system
about a common point, which usually turns out to be the center of gravity of a
molecule.
If a molecule has two or more orientations in space that are indistinguishable,
the molecule possesses symmetry. Two possible orientations for the hydrogen
molecule can be illustrated in Fig. 1-1 only by labeling the two equivalent
hydrogen atoms in this figure with prime and double prime marks. Actually, the
two hydrogens are indistinguishable, the two orientations are equivalent, and the
molecule has symmetry. The two orientations in Fig. 1-1 can be obtained by
rotation of the molecule through 180 about an axis through the center of and
perpendicular to the hydrogen-hydrogen bond axis. This rotation is referred to
as a symmetry operation,and the rotation axis is called a symmetry element. The
terms symmetry element and symmetry operation should not be confused or used
interchangeably. The symmetry element is the line, point, or plane about which
the symmetry operation is carried out. The operation can be defined only with
respect to the element, and the existence of the element can be shown only by
carrying out the operation.
A simple test can be performed to verify the presence of symmetry. If you
were to glance at a structure, turn your back and have someone perform a
symmetry operation, on once again examining the molecule, you would not be
able to determine that a change had been made. The symmetry operations of a
molecule form a group and, as a result, are amenable to group theoretical Figure 1-1 Equivalent
procedures. Mastery of the material in this chapter is essential to the understand- orientations for H2.
ing of most of the material in this book.
2 Chapter 1 Symmetry and the Point Groups
1-2 SYMMETRY ELEMENTS
Five types of symmetry elements will be considered for point symmetry: (1) the
center of symmetry (inversion center), (2) the identity, (3) the rotation axis, (4)
the mirror plane, and (5) the rotation-reflection axis. The symmetry operations
corresponding to these elements will be defined in the course of a further discussion
of each element.
The Center of Symmetry, or Inversion Center

A molecule is said to possess a center of symmetry or inversion center when
every atom in the molecule, if moved in a straight line through this center and
N _T~N an equal distance on the other side of the center, encounters a like atom. If oxygen
Inversion -B atom A of the hyponitrite ion (Fig. 1-2) is moved through the inversion center,
center it comes into coincidence with another oxygen atom, B. The same arguments
FIGURE 1-2 The center of must apply to atom B and also to both nitrogen atoms if the molecule is to
symmetry in the hyponitrite possess a center of symmetry. This operation corresponds to placing this center
ion. at the center of a coordinate system, and taking every atom with coordinates
(x, y, z) and changing its coordinates to (- x, -y, -z). At most one atom can
be at the center, and all other atoms in the molecule must exist in pairs. Neither
of the two structures indicated in Fig. 1-3 nor any other tetrahedral molecule
possesses an inversion center. Other molecules or ions that should be examined
CI H for the presence of an inversion center are 1,4-dioxane, tetracyanonickelate(II),
Br C1 trans-dichloroethylene, and trans-dichlorotetraammine cobalt(III). The molecule
C1
HCCl=CHBr (cis or trans) does not possess a center of symmetry because the
cIl CI ci Br C1 symmetry operation on bromine does not result in coincidence. All points in the
cl
molecule must be inverted in this operation, if the molecule as a whole is to
possess this symmetry element. The symbol used to indicate an inversion center
FIGURE 1-3 Absence of a
(center of symmetry ) is i.
center of symmetry in CCl 4
and HCCI-=CBrCI.
The Identity
In the identity operation, no change is made in the molecule. Obviously, the
result of this operation is not only to produce an equivalent orientation but an
identical one; i.e., even if similar atoms were labeled with prime or double prime
marks, or any other notation, no change would be detected. All molecules possess
this symmetry element, and it is indicated by the symbol E. This operation
probably seems trivial at present, but, as will be seen in Chapter 2, this operation
is required so that symmetry elements can be treated by group theory.
The Rotation Axis

If an imaginary axis can be constructed in a molecule, around which the molecule
can be rotated to produce an equivalent (i.e., indistinguishable from the original)
orientation, this molecule is said to possess a rotation axis. The symmetry element
previously discussed for the hydrogen molecule is a rotation axis. This element
is usually referred to as a proper rotation axis. It may be possible to carry out
several symmetry operations around a single rotation axis. If the molecule can
occupy n different equivalent positions about the axis, the axis is said to be of
order n. For example, consider the axis through the center of the boron atom in
BCI3 perpendicular to the plane of the molecule. Rotation, which by convention Cl Clu
is in a clockwise direction, about this axis two times through an angle of 120 BB
each time produces two equivalent orientations. Taken with the initial orientation, Cl"
we have the three different equivalent orientations, illustrated in Fig. 1-4. The
order, n, of this axis is three, for three rotations are needed to return to the
original position. The molecule is said to possess a threefold rotation axis,
indicated by the symbol C3 . Rotation of the molecule through 2 nE/n (i.e., 120 .13
Cl C1
when n = 3) produces equivalent orientations, and n operations produce the B
starting configuration referred to as the identity. The symbol C 32 is employed
to indicate a rotation of 240 around a C 3 axis. The C 32 operation is identical Cl,
to a counterclockwise rotation of 1200, which is indicated as C 3 -. It should be
clear that an axis through the center and perpendicular to the plane of a benzene
ring is a sixfold axis, C. Since n = 6, rotation by 60 (= 360 /6) six times 3 1
CI' Cl"
produces the six equivalent orientations. Further examination of the BCl 3
B
molecule indicates the lack of a center of symmetry and the presence of the three
additional twofold rotation axes, C 2 , illustrated in Fig. 1-5. One twofold axis, Cl
arbitrarily selected, is labeled C 2 ; the other two are labeled C 2 ' and C2 ". If n = 1,
FIGURE 1-4 The results of
the molecule must be rotated 3600 to produce an equivalent (and in this case clockwise rotations about the
identical) orientation. A molecule is said to possess no symmetry if no elements threefold rotation axis of BCl3.
other than the identity are present.
C1 Cl
B B
Cl' CI"1
C2 Rotation Cl" Cl
axis FIGURE 1-5 The results of
CI' CI' rotations about the three twofold
B B rotation axes in BC2. The 180
C11 CI' Cl1 C1l rotation on the left produces the
result on the right.
1C1 Cl"
B B
Cl" Cl' H-H4
Cl CI'
Rotation
A rotation axis of order n generates n operations: C , C. 2 , C 3 ,. C,-1, C,". axis
Furthermore, the operation C, 2 is equivalent to C2 , the operation C' 2 is FIGURE 1-6 The C, axis in
equivalent to C 3 , and C." is the identity. The highest-fold rotationaxis is referred H2.
to as the principal axis in a molecule. If all of the Cn axes are equivalent, any one
may be -chosen as the principal axis.
In Fig. 1-6, a rotation axis is illustrated for the H 2 molecule, for which n = o.
The C 2 rotation axes are not illustrated but are perpendicular to the bond axis
and centered between the two hydrogen atoms. There are an infinite number of
these twofold axes. It should also be obvious that benzene possesses six twofold
axes that lie in the plane of the molecule. Three pass through pairs of opposite F
C2 Rotation
carbon atoms, and the other three pass through the centers of C-C bonds. axis
The molecule CIF 3 is illustrated in Fig. 1-7. The geometry is basically a trigonal
bipyramid, with two lone pairs of electrons in the equatorial plane and with two
fluorines bent down toward the equatorial fluorine. This molecule has only one F
rotation axis, the C 2 axis shown in Fig. 1-7. The reader should verify the absence FIGURE 1-7 The symmetry
of symmetry along the axes indicated by dashed lines. axis in CIF 3.
For many purposes, it is convenient to locate a molecule and its symmetry

elements in a Cartesian coordinate system. A right-handed coordinate system
will be used. Movement of y to x is a positive rotation. The center of gravity of
the molecule is located at the center of the coordinate system. If there is only
one symmetry axis in the molecule, it is selected as the z-axis. If there is more
than one symmetry axis, the principal rotation axis in the molecule is selected
as the z-axis. If there is more than one highest-fold axis, the one connecting the
most atoms is selected. If the molecule is planar and if the z-axis lies in this plane
(as in the water molecule), the x-axis is chosen perpendicular to this plane. If the
molecule is planar and the z-axis is perpendicular to this plane, e.g., in
Cl H
=C , the x-axis is chosen to pass through the largest number of atoms.
H CI
In trans-dichloroethylene the x-axis passes through the two carbons.
The Mirror Plane or Plane of Symmetry

If a plane exists in a molecule that separates the molecule into two halves that
are mirror images of each other, the molecule possesses the symmetry element
of a mirror plane. This plane cannot lie outside the molecule but must pass
through it. Another way of describing this operation involves selecting a plane,
dropping a perpendicular from every atom in the molecule to the plane, and
placing the atom at the end of the line an equal distance to the opposite side of
the plane. If an equivalent configuration is obtained after this is done to all the
atoms, the plane selected is a mirror plane. Reflection through the mirror plane
is indicated by Fig. 1-8. A linear molecule possesses an infinite number of mirror
planes.
reflect
FIGURE 1-8 Operation of a r O-c0
mirror plane (a) reflection on H H' H' H
H20.
Mirror plane
Often rotation axes lie in a mirror plane (see Fig. 1-9, for example), but
there are examples in which this is not the case. The tetrahedral molecule POBr 2 Cl,
illustrated in Fig. 1-10, is an example of a molecule that contains no rotation
axis but does contain a mirror plane. The atoms P, Cl, and 0 lie in the mirror
plane. In general, the presence of a mirror plane is denoted by the symbol o. In
those molecules that contain more than one mirror plane (e.g., BCI3 ), the
horizontal plane a. is taken as the one perpendicular to the principal (highest-fold
rotation) axis. In Fig. 1-9, the plane of the paper is u, and there are then three
vertical planes a,, (two others, similar to the one illustrated, each containing the
boron atom and one chlorine atom) perpendicular to Uh.
When a molecule is located in a coordinate system, the-i-axis always lies in
the vertical plane(s).
Some molecules have mirror planes containing the principal axis but none
of the perpendicular C2 axes. These planes bisect the angle between two of the
C 2 axes (in the xy-plane); they are referred to as dihedralplanesand are abbreviated
as ad. Two ad planes are illustrated in Fig. 1-14, vide inp-a.* In some molecules
there is more than one set of mirror planes containing the highest-fold axes; e.g.,
in PtCl4 2 -, one set consists of the xz- and yz-planes (z being the fourfold axis,
and the Cl-Pt-Cl bond axes being x and y), and the other set bisects the angle
between the x- and y-axes. The former set, including the chlorine atoms, is called
a,, by convention, and the latter set is called ad. The diagonal plane is always
taken as one that bisects x, y, z, or twofold axes in the molecule.
The Rotation-Reflection Axis; Improper Rotations C2

Axis
This operation involves rotation about an axis followed by reflection through a
FIGURE 1-9 A mirror plane
mirror plane that is perpendicular to the rotation axis, or vice versa (i.e.,
in BC3 .
rotation-reflection is equivalent to reflection-rotation). When the result of the
two operations produces an equivalent structure, the molecule is said to possess
a rotation-reflection axis. This operation is referred to as an improper rotation,
and the rotation-reflection axis is often called an alternating axis. The symbol S
is used to indicate this symmetry element. The subscript n in S, indicates rotation
(clockwise by our convention) through 2nr/n.
Obviously, if an axis C, exists and there is a a perpendicular to it, C,, will
also be an S.. Now, we shall consider a case in which S, exists when neither C,
Br
or the mirror plane perpendicular to it exists separately. In the staggered form
of ethane, Fig. 1-11, the C-C bond defines a C3 axis, but there is no perpendicular
mirror plane. However, if we rotate the molecule 60 and then reflect it through
a plane perpendicular to the C-C bond, we have an equivalent configuration.
Consequently, an S, axis exists and, clearly, there is no C,. FIGURE 1-10 The mirror
The dotted line in Fig. 1-12 indicates the S2 element in the molecule plane in POBr 2 CI.
trans-dichloroethylene. The subscript two indicates clockwise rotation through
180". S2 is equivalent to i and, by convention, is usually called i.
H H H- C
C6 a
HHH
HC - H
H H/
FIGURE 1-11 Improper rotation axis in the staggered form of ethane.
A particular orientation of the molecule methane is illustrated in Fig. 1-13.

Open circles or squares represent hydrogen atoms in a plane parallel to but above
the plane of the paper, and the solid squares or circles are those below the plane
* We will use this term to indicate that a statement will be treated more fully later in the text.
H CI of the paper. The plane of the paper is the reflection plane, and it contains the
\ / carbon atom. The C4 operation is straightforward. The operation of reflection,
- - - C-C- -S 2 a, moves the hydrogens below the plane to above the plane and vice versa. This
/ \
Cl H is indicated by changing the solid squares to open squares and the open circles
FIGURE 1-12 Rotation- to solid circles. However, since all four hydrogens are identical, the initial and
reflection axis of symmetry. final orientations are equivalent. The molecule contains a fourfold rotation-
reflection axis, abbreviated S4 . This operation can be repeated three more times,
or four times in all. These four operations are indicated by the symbols S , S 4 ,
2
S4 3 , and S44. The reader should convince himself or herself that S4 is equivalent
to a C 2 operation on this axis. It should also be mentioned that the molecule
possesses two other rotation-reflection axes, and these symmetry elements are
3
abbreviated by the symbols S' and S". The operation S4 is equivalent to a
counterclockwise rotation of 900 followed by reflection. This is often indicated
as S 4 -.
FIGURE 1-13 Effect of the

operations C4 and a
C4 a
perpendicular to C4 on the * 00
hydrogens of CH 4 . (Open o Co C) oC
symbols above the page; M 0 0
solid symbols below.)
Next we shall consider some differences in improper rotation axes of even

and odd order. With n even, S.' generates the set S Sf, 2, Sn, ... , S.". This is
equivalent to C,,a, C,7a , C 33, ... , C,"a". (Note:
2
In carrying out C,"o", one
reflects first and then rotates, while in carrying out U"C,", one rotates and then
reflects.) We have the relation a' = a when m is odd, and a-' = E when m is even.
The latter leads to the identities S," = C,"E = E and S.' = C,,'.
Notice that the existence of an S,, axis of even order always requires a C.2,
since S,2 = Cn2 a2 = Cz 12 -
Consider the set S6, S6 2 , 3, 4 S66 by carrying out the operations
described below on the staggered ethane example.
S6 cannot be written any other way.

2 2
S6 2 = C6 U = C 3
S63 = S2 = i
(An S2 axis always equals i.)
2
S6 = C3
Se6 cannot be written any other way.

6
S6 = E
The complete set would normally be written:
So C3 i C3 S65 E
When n is odd, S.'"generates the set S,, S,21 S3, Sn 4, ... , S2n". An odd-order
S,, requires that C, and a a perpendicular to it exist. Note that the operation S,"
when n is odd is equivalent to C,,"a" = C,"a = a. (Compare this with S," when
n is even.)
For example, consider S3 . First one would reflect, producing a configuration
that is equivalent to the starting configuration (because of the existence of 0 in
the molecule). Then one rotates by C3 to give the configuration corresponding to
S3 1. Since S 31 is a symmetry operation for the molecule, the configuration after
the C3 rotation must be equivalent to the starting one, and therefore C3 exists.
In general, for any S. (with n odd), C is also a symmetry operation. For practice,
consider the S,' operations:
S, = C, and then a
S5 2 = C5 2
3
S53 = C and then a
4
S = C5
-
S5 -
S5 6
= C5
S = C, 2 and then a
8 3
S5 -C 5
S" = C.' and then a

S,10 = E
1
S = SS, etc.
The boldface indicates S5 operations that cannot be written as a single

operation. Thus, S, with n odd generates 2n operations. Table 1-1 summarizes
the key aspects of our discussion of symmetry operations and elements.
TABLE 1-1. A Summary of Symmetry Elements and Symmetry Operations

Symmetry Operation Symmetry Element Symbol Examples
identity E all molecules
reflection plane a H 20, BF3 (planar)
Cl,
inversion point (center of i C B-B , Cl
symmetry) Cl Cl
proper rotation axis C. (n-order) NH3, H20
improper' rotation axis and plane S,(n-order) ethane, ferrocene
(rotation by 27r/n followed (staggered structures)
by reflection in plane
perpendicular to axis)
aFerrocene is staggered and possesses an S10 improper rotation axis.

1-3 POINT GROUPS
A point group is a collection of all the symmetry operations that can be carried
out on a molecule belonging to this group. It is possible to classify any given
molecule into one of the point groups. We shall first examine molecules belonging
to some of the very common, simple point groups for the symmetry elements
they possess. This will be followed by a general set of rules for assigning molecules
to the appropriate point groups.
A molecule in the point group C,, has only one symmetry element, an n-fold
rotation axis. (Of course, all molecules possess the element E, so this will be
assumed in subsequent discussion.) A molecule such as trans-dichloroethylene,
which has a horizontal mirror plane perpendicular to its C, rotation axis, belongs
to the point group C,,. The point group C,,, includes molecules like water and
sulfuryl chloride, which have n vertical mirror planes containing the rotation
axis, but no horizontal mirror plane (the horizontal plane by definition must be
perpendicular to the highest-fold rotation axis). The C 2 , molecules, H 2 0 and
S0 2C12 , contain only one rotation axis and two a, planes. The molecule is
assigned to the point group that contains all of the symmetry elements in the
molecule. For example, H 20, which has a C2 axis and two vertical planes, is
assigned to the higher-order (more symmetry elements) point group C 2, rather
than to C 2 .
The symbol D, is used for point groups that have, in addition to a C, axis,
n C 2 axes perpendicular to it. Therefore, the D. point group has greater symmetry
(i.e., more symmetry operations) than the C. group. A D, molecule that also has
a horizontal mirror plane perpendicular to the C, axis belongs to the point group
D,,, and as a consequence will also have n vertical mirror planes. The addition
of a horizontal mirror plane to the point group C,,, necessarily implies the presence
of n C 2 axes in the horizontal plane, and the result is the point group D... BC13
is an example of a molecule belonging to the D 3h point group. In the D, point
groups, the a, planes are perpendicular to the principal axis and contain all the
C 2 axes. Each a, plane contains the principal axis and one of the C 2 axes.
C'£
H -
C2 - C2
FIGURE 1-14 The C2-axes and two Ud planes in allene.

1-3 Point Groups 9
TABLE 1-2. Symmetry Elements in Some Common Point Groups
Point Group Symmetry Elementsa Examples
C, no symmetry SiBrCIFI
C2 one C2 axis H 202
C.h one n-fold axis and a horizontal plane u, trans-C2 HCl2 (C2,,)
which must be perpendicular to the n-fold axis
C2 one C 2 axis and two a. planes H2O, SO2Cl2, SiCl2Br2
C3 one C3 axis and three a, planes NH3, CH 3 Cl, POCI,
D2 h three C2 axes all I, two u, planes, one a, plane, N 20 4 (planar)

and a center of symmetry
D3h one C3 , three C2 axes I to C3, three a, BCl 3

planes, and one a,
D 2d three C2 axes, two ad planes, and one S4 H2C=C=CH2

(coincident with one C2 )
Td three C2 axes I to each other, four C3 , six a, CH 4, SiCI4

and three S4 containing C 2
'All point groups possess the identity element, E.
D, molecules may also have a, planes that contain the principal (C.) axis
but none of the perpendicular C 2 axes. As mentioned previously, these dihedral
planes, ad, bisect the angle between two of the C2 axes. The notation for a D,
molecule containing this symmetry element is Dfd. Such a molecule will contain
an n-fold axis, n twofold axes perpendicular to C., and in addition n (vertical)
planes of symmetry bisecting the angles between two twofold axes and containing
the n-fold axis. Allene, H 2C=C=CH 2 , is an example of a molecule belonging
to the D 2d point group. Some of its symmetry elements are illustrated in Fig.
1-14. The two hydrogens on one carbon are in a a, plane perpendicular to the
a, plane containing the two hydrogens on the other carbon. The C 2 axes (the
principal axis being labeled C 2 , the others C 2 ' and C2 ") and the dihedral planes
containing the principal axis are indicated. The two other C 2 axes, C 2 ' and C 2",
which are perpendicular to the principal C 2 axis, form 450 angles with the two
dihedral planes. (Study the figure carefully to see this. Constructing a model may
help.)
Table 1-2 contains the symmetry elements and examples of some of the
more common point groups. The reader should examine the examples for the
presence of the symmetry elements required for each point group and the absence
of others. Remember that the symbol C indicates that the molecule has only one
rotation axis. The symbol D, indicates n C 2 axes in addition to the n-fold axis;
e.g., the point group D4 contains a C 4 axis and four C 2 axes.
H 0-H A more exhaustive compilation of the important point groups is contained

\ / in Herzberg.m) In addition to the point groups listed in Table 1-2, the 0 point
O-B
group which includes molecules whose structures are perfect octahedra
0
(SF6 , PC 6 ) is very common.
H
It is important to realize that although CHCl 3 has a tetrahedral geometry,
FIGURE 1-15 A structure it does not have tetrahedral symmetry, so it belongs to the point group C3 v and
with C3h symmetry. not T. A tetragonal complex, trans-dichlorotetraammine cobalt(III) ion, belongs
to the point group D4 h (ignoring the hydrogens) and not 0 Phosphorus
pentachloride belongs to the point group D3h, and not to C3h, for it has three
C 2 axes perpendicular to the C 3 axis. The structure of monomeric boric acid
(assumed to be rigid), illustrated in Fig. 1-15, is an example of C 3h symmetry. It
has a threefold axis and a a, plane but does not have the three C 2 axes or 0,
planes necessary for the D3h point group.
It is helpful but not always possible to classify additional structures by
memorizing the above examples and using analogy as the basis for classification
instead of searching for all the possible symmetry elements. The following sequence
of steps has been proposedm" for classifying molecules into point groups and is
a more reliable procedure than analogy.
1. Determine whether or not the molecule belongs to one of the special

0
point groups, C., D , , h, or T. The group I, contains the regular
dodecahedron and regular icosahedron. Only linear molecules belong to C, or
Dh.
2. If the molecule does not belong to any of the special groups, look for a
proper rotation axis. If any are found, proceed to step (3); if not, look for a center
of symmetry, i, or a mirror plane, a. If the element i is present, the molecule
belongs to the point group Cj; if a mirror is present, the molecule belongs to the
point group C. If no symmetry other than E is present, the molecule belongs to C 1.
3. Locate the principal axis, C,,. See if a rotation-reflection axis S2, exists
that is coincident with the principal axis. If this element exists and there are no
other elements except possibly i, the molecule belongs to one of the S. point
groups (where n is even). If other elements are present or if the S2n element is
absent, proceed to step (4).
4. Look for a set of n twofold axes lying in the plane perpendicular to C..
If this set is found, the molecule belongs to one of the groups D,, D,1, or Dnd-
Proceed to step (5). If not, the molecule must belong to either Cn, C,, or C,
Proceed to step (6) and skip (5).
5. By virtue of having arrived at this step, the molecule must be assigned
to D, Dn, or D, If the molecule contains the symmetry element a,, it belongs
to Dfl. If this element is not present, look for a set of n ad's, the presence of
which enables assignment of the molecule to Dad. If o- and o-, are both absent,
the molecule belongs to D..
6. By virtue of having arrived at this step, the molecule must be assigned to
C,1 Ch,, or Cn, . If the molecule contains ah, the point group is Cn1. If ch is absent,
look for a set of n a,'s, which places the molecule in Cn,. If neither a, nor 6, is
present, the molecule belongs to the point group C..
The following flow chart provides a systematic way to approach the

classification of molecules by point groups:
1-4 Space Symmetry 11
"Special Groups"
S,, or S2 .and i only,

collinear with unique ur
highest order C.
Unit
cell
to
1-4 SPACE SYMMETRY
In describing the symmetry of an arrangement of atoms in a crystal (essential to

FIGURE 1-16 A glide plane.
x-ray crystallography), two additional symmetry elements are required: (1) the
(The tails of these commas
glide plane and (2) the screw axis, corresponding to mixtures of point group
point into the page.)
operations and translation. A space group is a collection of symmetry operations
referring to three-dimensional space. A glide plane is illustrated in Fig. 1-16. The
unit (represented by a comma) is translated a fraction (in this case '/2) of a unit
cell dimension and is reflected. (A unit cell is the smallest repeating unit of the 6-
crystal.) The comma must lie in a plane perpendicular to the plane of the papet x
and all of the comma tails must point into the paper. A screw axis is illustrated
in Fig. 1-17. Each position, as one proceeds from top to bottom of the axis, 0- ~.0 X
represents rotation by 120 along with translation parallel to the axis. This is a
6-
,x
threefold screw axis. Rotations by 90 , 60 , and other angles are found in other
structures. There are 230 possible space groups. These are of importance in x-ray
crystallography, and will be covered in more detail in that chapter. FIGURE 1-17 A screw axis.
1-5 SOME DEFINITIONS AND APPLICATIONS OF SYMMETRY

CONSIDERATIONS
Products or Combinations of Symmetry Operations

The product of any two symmetry operations, defined as their consecutive
application, must be a symmetry operation. Hence, the product of the C2 and
a,' operations on a molecule with C2,, symmetry (e.g., H 2 O) is a,. (Recall that
the C2 axis is the z-axis, the plane of the molecule is yz, and a,' is the yz-plane.)
This can be written as:
C2 x a,' = a. or
C2av' =a6
Instead of the term product, the term combination is a better description of the
above operations. The order in which the operations are written (i.e., from left
to right) is the reverse of the order in which they are applied. For the above
example, the o,,' operation is carried out first and is followed by C 2 to produce
the same result as a,. The final result often depends on the order in which the
operations are carried out, but, in some cases, it does not. When the result is
independent of the order (e.g., C2 Uc' = U,'C2), the two symmetry elements C 2 and
a,' are said to commute. In a D 3h molecule (see, e.g., Fig. 1-4), the two operations
C 3 and a, do not commute; i.e.,
C 3a, # aC 3
Equivalent Symmetry Elements and Equivalent Atoms

If a symmetry element A can be moved into the element B by an operation
corresponding to the element X, then A and B are said to be equivalent. If we
define X -1 as the reverse operation of X (e.g., a counterclockwise rotation instead
of a clockwise one), then X-' will take B back into A. Furthermore, if A can be
carried into C, then there must be a symmetry operation (or a sequence of
symmetry operations) that carries B into C, since B can be carried into A. The
elements A, B, and C are said to form an equivalent set.
Any set of symmetry elements chosen so that any member can be transformed
into each and every other member of the set by application of some symmetry
operation is said to be a set of equivalent symmetry elements. This collection of
CI C elements is said to constitute a class. To make this discussion clearer, consider
the planar molecule PtCl4 2 - illustrated in Fig. 1-18. The C2' and C 2 axes form
an equivalent set, as do the C2 " and C2 "'. However, C2' is not equivalent to C2 ",
CL-Pt ci- C because the molecule possesses no symmetry operation that takes one into the
other.
ClC2 Equivalent atoms in a molecule are defined as those that may be interchanged
with one another by a symmetry operation that the molecule possesses.
2
Accordingly, all of the chlorines in PtCl4 - (Fig. 1-18) and all of the hydrogens
C2 in methane, benzene, cyclopropane, or ethane are equivalent. The fluorines in
gaseous PF, (trigonal bipyramidal structure) are not all equivalent, but form two
FIGURE 1-18 Equivalent
symmetry elements and sets of equivalent atoms, one containing the three equatorial fluorines and one
atoms in the molecule PtCl- containing the two axial fluorine atoms. These considerations are very important
1-5 Some Definitions and Applications of Symmetry Considerations 13
when the topic of nuclear magnetic resonance is discussed, for, under favorable
conditions, non-equivalent atoms give rise to separate peaks in the spectrum.
Optical Activity
If the mirror image of a molecule cannot be superimposed on the original, the
molecule is optically active; if it can be superimposed, the molecule is optically
inactive. In using this criterion, the mirror is understood to be external to the
whole molecule, and reflection through the mirror gives an image of the whole
molecule. With complicated molecules, the visualization of superimposability is
difficult. Accordingly, it is to our advantage to have a symmetry basis for
establishing the existence of optically active isomers. Any molecule that has no
improper rotation axis is said to be dissymmetric, and optically active molecules
must be dissymmetric. One often hears the incomplete statement that in order for
optical isomerism to exist, the molecule must lack a plane or center of symmetry.
7'\ CH3
Since S, = a (Si is a rotation by 360 followed by a reflection) and S2 = i, if the
molecule lacks an improper rotation axis, both i and a must be lacking. To show
the incompleteness of the earlier statement, we need to find a molecule that has
neither a nor i, but does contain an S, axis and is not optically active. Such a
molecule is 1,3,5,7-tetramethylcyclooctatetraene, shown in Fig. 1-19. This mol- CH3
ecule does not have a plane or center of symmetry. However, since it has an S4 FIGURE 1-19 Structure of
axis, it is not optically active. 1,3,5,7-tetramethylcyclo-
A dissymmetric molecule differs from an asymmetric one,for the latter type is octatetraene.
completely lacking in symmetry. The molecule trans-1,2-dichlorocyclopropane,
shown in Fig. 1-20, is dissymmetric (there is no S, axis) and hence optically
active, but it is not asymmetric, for it possesses a C2 axis.
Many molecules can exist in some conformation that is optically active. H H
However, if rotation of the molecule about a bond produces a conformation with
an improper axis, the molecule will not be optically active. If the conformation C
is frozen in a form that does not possess an improper axis, optical activity could
result. H, C1
In summary, then, we can state that if a molecule possesses only C., it is C C
dissymmetric and optically active. If n = 1, the molecule is asymmetric as well
as dissymmetric; and if n > 1, the molecule is dissymmetric. If a molecule possesses CI H
S. with any n, it cannot be optically active. C2
FIGURE 1-20
Dipole Moments Trans-1,2-
dichlorocyclopropane.
Molecules possess a center of gravity of positive charge, which is determined by
the nuclear positions. When the center of gravity of the negative charge from the
electrons is at some other point, the molecule has a dipole moment, which is
related to the magnitude of the charge times the distance between the centers.
The dipole moment is a vector property; that is, it has both a magnitude and a
direction. For a fixed geometry (i.e., a non-vibrating molecule) the dipole moment
is a non-fluctuating property of the molecule; as a result it, like the total energy,
must remain unchanged by the operation of every symmetry element of the
molecule. In order for this to occur, the dipole moment vector must be coincident
with each of the symmetry elements.
There are several obvious consequences of these principles. Molecules that
have a center of symmetry cannot possess a dipole moment, because the vector
cannot be a point, and any vector would be changed by inversion through the
center. Molecules with more than one C,, axis cannot have a dipole moment, for
a dipole moment vector could not be coincident with more than one axis. Thus,
only the following types of molecules may have dipole moments: those with one
C,. (n > 1), those with one - and no C., those with a C, and symmetry planes
that include C,,, and those that have no symmetry. In all cases, where the molecule
has symmetry, the direction of the dipole vector is determined; for it must lie in
all of the symmetry elements that the molecule possesses.
REFERENCES 1. G. Herzberg, "Infrared and Raman Spectra," p. 11, Van Nostrand, New York, 1945.
2. F. A. Cotton, "Chemical Applications of Group Theory," p. 54, Wiley-Interscience,
CITED New York, 3rd ed., 1990.
ADDITIONAL M. Tinkham, "Group Theory and Quantum Mechanics," McGraw-Hill, New York, 1964.
F. A. Cotton, "Chemical Applications of Group Theory," 3rd ed., Wiley-Interscience,
READING New York, 1990.
M. Orchin and H. H. Jaff6, J. Chem. Educ., 47, 246, 372, 510 (1970).
M. Orchin and H. H. Jaffe, "Symmetry, Orbitals, and Spectra," Wiley-Interscience, 1971.
C. D. H. Chisholm, "Group Theoretical Techniques in Quantum Chemistry," Academic
Press, New York, 1976.
A. Vincent, "Molecular Symmetry and Group Theory," Wiley, New York, 1977.
For space groups, see: J. D. Donaldson and S. D. Ross, "Symmetry and Stereochemistry,"
Wiley, 1972.
EXERCISES 1. Classify the following molecules in the appropriate point groups and, for (c)through
(k), indicate all symmetry elements except E.
a. CoCl 2 -
2
b. Ni(CN),
c. cis-CoCl 4 (NH 3)2 (ignore the hydrogen atoms)
d. CH 2 (chair form)
e. Si(CH 3)3 -A -B (with A and B trans in a trigonal bipyramid)
f. PF 3
g. (CH 3)2B H B(CH 3)2
H
h. Cl-I-Cl-
i. planar cis-PdCl 2B2 (B = base)

Exercises 15
j. planar trans-PdC 2 B2 (B = base)
k. staggered configuration for C 2 H,
1. ferrocene (staggered)
2. a. How does a D 2 d complex for formula MCI4 2 - differ from a Td complex (i.e., which
symmetry elements differ)?
b. What important symmetry element is absent in the PF3 molecule that is present
in the D 3h point group?
2
c. If each Ni-C-N bond in planar Ni(CN)4 - was not linear but was bent
C
(Ni N), to what point group would this ion belong? What essential symmetry
element present in D4, is missing in C4h?
d. Indicate which operation is equivalent to the following products for the T, point
group
(1) S4 x S4 =?
(2) C 3 x C 2 =?
(3) ad x C 2 =?
3. Does POCI 2 Br possess a rotation-reflection axis coincident with the P-O bond axis?
4. a. Rotate the molecule in Fig. 1-14, 1800 around the S4 axis, and draw the resulting
structure.
b. Locate the mirror plane of the S4 operation and indicate on the structure resulting
from part (a) where all the atoms are located after reflection through the plane.
Use script letters to indicate the final location of each atom.
c. Is S2 a symmetry operation for the molecule?
5. To which point group does the molecule triethylenediamine (also called 1,4-di-
azabicyclo[2.2.2]octane, or dabco) belong?
H2 C
H2C N CH2
CH2
H 2CN CH2
6. Assign PtCl4 2 - to a point group and locate one symmetry element in each class of
this point group.
7. Give the point group for each of the following and indicate whether the entity could
have a dipole moment:
a. three-bladed propeller
b. hexahelicene
Above
Below
Oro
00
c. 2,2-cyclophane
d. hexadentate ligand on Fe 2 +
e. Co(C 5HNO)6 2+
(CHNO is C-C)N-
C-C
Q =Co
D =o
Exercises 17
8. Is the following compound optically active?
CH3 H
H CH 3
H--- H
H H
N' +
H .H
H H
H CH 3
CH3 H
9. Cr(en) 3 3 * (where en = ethylenediamine) belongs to the D, point group. Is it optically

active? Why?
10. Determine the point groups of the following molecules, and find the requested
equivalent atoms within these molecules.
a. Cl Cl Which protons, if any, are equivalent?
H3 Hi H2
Cl
b. Cl Which protons are equivalent?

H 8 ,DH 5
H7 H,
Fe
H Hi
Cl
c. H6 Cl Which protons are equivalent?
H H
4 gH
2
Cl H
d. Cis" C1,, C" Which chlorides are equivalent?

Cu Cu
Cl7 C14 C12
11. Locate the a, planes in CH,. Are these equivalent to a,? Why?
Group Theory and
the Character
Tables
2-1 INTRODUCTION
Group theory is a topic in abstract algebra that can be applied to certain systems
if they meet specific requirements. There are many systems of interest to chemists
that can be treated by these techniques, and by the end of this chapter you will
only have begun to gain an appreciation for the power of this method. We shall
use these concepts frequently throughout the rest of the book. We shall be involved
in some abstract and seemingly irrelevant concepts in the course of our
development of this topic. The reader is encouraged to persevere, for the rewards
are many.
There are two common major types of applications of group theory.
a. It can be used in the generation of symmetry combinations. If you are
given some basis set, that is, a set of orbitals or mathematical functions pertaining
to a molecule, you can use group theory to construct linear combinations of the
things in the basis set that reflect the symmetry of the molecule. For example, if
you are given a set of atomic orbitals for a molecule, you can use group theory
to assist you in mapping out the shape of the molecular orbitals for the given
molecule. There are many areas where such an application can be made, and a
list of a few include our ability to:
1. determine hybrid orbitals used in bonding the atoms in the molecule.
2. determine which atomic orbitals can contribute to the various molecular
orbitals in a molecule
3. determine the number and symmetries of molecular vibrations
4. predict how the degeneracies of the d-orbitals are removed by crystal
fields of various symmetries
5. predict and generate the spin functions to be used in the Hamiltonian
for epr, nqr, and M6ssbauer spectroscopy
6. construct symmetry adapted combinations of nuclear spin functions to
work out nmr equations.
b. It can be used to ascertain which quantum mechanical integrals are zero;

i.e., isjf* op 0 dr equal to zero? Such integrals are important to:
1. determine the allowedness of electronic transitions
2. determine the activity of infrared and Raman vibrations
18 3. determine the allowedness of any given transition in nmr, esr, etc.
2-2 RULES FOR ELEMENTS THAT CONSTITUTE A GROUP
The symmetry operations can be treated with group theory, and we shall use
them to illustrate the principles. In order for any set of elements to form a
mathematical group the following conditions must be satisfied.
1. The combination (Chapter 1; often referred to as the "product") of any
two elements, and the square of each element, must produce an element of the
group.
This combination is not the same thing as a product in arithmetic. We are
referring to the consecutive operation of elements. It is a simple matter to square
and take all possible combinations of the elements of the C 2, point group to
illustrate the point that another element is obtained. The reader is encouraged
to do so, using H 2 0 as an example. Unlike multiplication in ordinary arithmetic,
we must worry about the order of the combination, i.e., C 2 x a, or o- x C 2 -
2. One operation of the group must commute with all others and leave them
unchanged. This is the identity element.
Eo2 = 6 2 E = 2
3. The associative law of multiplication must hold, i.e.,
(XY)Z = X(YZ)
The result of combining three elements must be the same if the first element
is combined with the product of the second two (i.e., o,(C2 ,') or if the product
of the first two is combined with the last element (i.e., (aC2)a,')-
4. Every element must have a reciprocal that is also an element of the
group. This requires that for every symmetry operation there be another operation
that will undo what the first operation did. For any mirror plane, the inverse is
the identical mirror plane, e.g., a x a = E. For a proper rotation C,', the inverse
is C" -"; C.'" x C,"-' = E. If A has the reciprocal element B, then AB = BA = E.
If B is the reciprocal of A, then A is the reciprocal of B. In general, the reciprocal
of A can be written as A1.
C3 has the reciprocal element C 3 2 :
C 3C 2
= C 32C 3 =E
We shall illustrate these rules in an abstract way in the next section.
2-3 GROUP MULTIPLICATION TABLES
Properties of the Multiplication Tables

If we have a complete and non-redundant list of all the elements in a finite group,
and we know all the products, then the group is completely and uniquely defined.
This information can be summarized with a group multiplication table. This table
is simply an array in which each column and row is headed by an element. The
matrix format is employed here because it is a convenient way to insure that we
have taken all possible permutations of products of the elements. The number
20 Chapter 2 Group Theory and the Character Tables
of elements in the group is referred to as the order of the group, h. The group
multiplication table will consist of h rows and h columns. Certain groups have
an infinite number of elements, e.g., H 2 belongs to Dc,.
In order to illustrate simply the rules of group theory discussed earlier, we
will consider a group of order three containing the abstract elements A and B
as well as the identity E. We shall determine what requirements these elements
must satisfy (i.e., how their combinations must be defined) in order for these
elements to constitute a group. Each matrix element is the result of the product
of the column element times the row element:
E AB
E
A
B
Since multiplication is generally not commutative, we must adhere to a consistent

order for multiplication. The convention is to carry out the operations in the
order, column element times row element; i.e., when we write a product RC, where
C is column and R is row, we take the column element first, followed by the row.
Each of the original h elements in the group must appear once and only
once in each row and each column of the group multiplication table, and no two
rows or columns may be alike. The following argument proves this. For a group
of h different elements E, AI. .. A,. ... A, 1, the elements of the nth row (An)
are A.E, A.A. ... AfAh_,. If any two products AnAi and AnA 1 were the same,
1
left multiplication by the inverse of A, would give A.- AnAj = Ai and
An- AnAj = Ai respectively, and thus require Ai and A1 to be the same. This
conclusion is contrary to our original assumption that all elements of the group
were different. Thus, all the products and all the h elements of the row must be
different. Each of the original h elements in the group must thus appear once
and only once in the row. A similar argument can be used for the columns. We
can illustrate the requirements for the general, unspecified set of elements E, A,
and B described above to constitute a group, and show how these rules define
the multiplication table.
Since E is the identity, it is a simple matter to indicate the matrix elements
for products involving E and write (REMEMBER column first, then row
multiplication):
E A B
E E A B
A A
B B
There are only two ways to complete this table; these involve defining either
AA = E or AA = B. If AA = E, then according to rule 4, A is its own reciprocal,
so BB = E. (A cannot be reciprocal to both B and A unless A = B, so B must
be the reciprocal of B.) Then, if BA = A or B, we would not be able to complete
the table without repeating an element in a row or column; i.e., BA = A makes
A repeat in a column, while BA = B repeats B in the row. However, if we define

AA = B, the following table results:
E A B
E E A B
A A B E
B B E A
If AA = B, we know that A must have a reciprocal, so AB = BA = E.

We can consider this entire group to be generated by taking an element and
all its powers, e.g., A, A 2 ( = B), and A 3(= E). Such a group is a cyclic group. For
such a group, all multiplications must commute.
Thus, by adhering to the conditions specified for a collection of elements to
constitute a group, we have been able to define our elements so as to construct
a group multiplication table that specifies the result of all possible combinations
of elements.
The very same procedures and rules can be used to construct group
multiplication tables of higher order. When a table of order four is made, two
possibilities result, which we shall label G,' and G, 2 :
2
G| E A B C G E A B C
E E A B C E E A
A A B C E A A E
B B C E A B B C H H'
C C E A B C C B
The superscripts arbitrarily number the two possibilities.

We can make these procedures more specific by considering a system in
which the elements are symmetry operations. In so doing, our definition of the H
products will not seem so arbitrary, for we know physically what the results of
the products of the symmetry operations are. We shall proceed by considering (Tv
the C 3 v ammonia molecule, whose point group contains the symmetry operations FIGURE 2-1 The a planes
E, 2C 3, and 3a, The three a, planes are labeled as in Fig. 2-1, and the group (.L to page) in the C3,
multiplication table is given as Table 2-1. The reader is encouraged to carry out ammonia molecule.
TABLE 2-1. Group Multiplication Table for the C3v Point

Group
and verify the combinations given in this table. All of the points made in
conjunction with the discussion of the group A, B, E should also be verified.
The group multiplication table (Table 2-1) contains sub-groups, which are
smaller groups of elements in the point group that satisfy all of the requirements
for constituting a group. E is such a sub-group, as is the collection of operations
E, C3 , C,. The order of any sub-group must be an integral divisor of the order,
h, of the full group.
Similarity Transforms
A second way in which the symmetry operations of a group may be subdivided
into smaller sets is by application of similarity transforms. If A and X are two
operations of a group, then X -'AX will be equal to some operation of the group,
say B:
B=X-'AX
The operation B is said to be the similarity transform of A by X. In the general

definition of a similarity tranform, X does not necessarily represent an operation
of the group. However, when taking the similarity transform of A in order for
the result B to be some other operation of the group, X must be an operation of
the group. We then say that A and B are conjugate. There are three important
properties of conjugate operations:
1. Every operation is conjugate with itself. If we select one operation, say

A, it must be possible to find at least one operation X such that A = X-'AX.
Proof:
Left multiply by A -':
A-A = E = A- 'X-'AX =(XA)-'AX
(The reciprocal of the product of two or more operations is equal to the product
of the reciprocals in reverse order; see Cotton,m) " Chemical Applications of Group
Theory," if you desire to see the proof.) By definition, E must also equal
(AX) -(AX).
Both equations can be true only if A and X commute. Thus, the operation
X may always be E, but it may also be any other operation that commutes with
the selected operation A.
2. If A is conjugate with B, then B is conjugate with A; i.e., if A =X -'BX,
then there must be some operation in the group, Y, such that B = Y-'A Y Note
that X equals Y -' and vice versa.
3. If A is conjugate with both B and C, then B and C are conjugate with
each other.
The use and meaning of the similarity transform will be made more concrete
as we employ it in subsequent discussion in this chapter. For the moment, it may
help to point out that a similarity transform can be used to change the coordinate
system selected to describe a problem. A right-handed coordinate system is given

by:
zY
or any rotation of these axes. A left-handed coordinate system is given by:
VY
The two coordinate systems are seen to be related to each other by a mirror
plane a, that contains the Z-axis and bisects the X- and Y-axes.
Now, consider a C,1 rotation about the Z-axis. A point with coordinates
(a, b) (the X coordinate is always given first) in a right-handed system goes to
(b, -a) upon a 900 clockwise rotation, but to (- b, a) with a 270 rotation.
In a left-handed system C,' takes (a, b) to (- b, a), while C, 3 yields (b, -a).
(Construct a figure if needed.) Thus, the roles of C4 and C, 3 are interchanged
by changing the coordinate system. We can say that there is a similarity transform
of C 3 to C4 by ad; i.e.,
C4 =d I C4ad
or, since ad-' = ad (note that adad- = E but adad = E so ad = ad-1)
C 4 =dC 4 ad
A A A A
B4 B, B3 B2 B2 B1 B3 B4
B3 B2 B4 B1 B3 B4 B2 B,
C4
Classes of Elements
A complete set of elements that are conjugate to one another is called a class of
the group. The order of each class must be an integral divisor of the order of the
group. To determine the classes, begin with one element, say A, and work out
all of its similarity transforms with all the elements of the group, including itself.
Then find an element that is not one of those conjugate with A, and determine
all of its transforms. Repeat this procedure until all of the elements in the group
have been divided into classes. E will always constitute a class by itself of order
I,E'EE= E.
We shall demonstrate this procedure by working out all of the classes of

elements in the group:
G6 1 E A B C D F
E E A B C D F
A A E D F B C
B B F E D C A
C C D F E A B
D D C A B F E
F F B C A E D
First of all, E is a class by itself. Next, work out all of the similarity transforms
of A. These would be EAE,A -'AA, B- AB,C -'AC,D 'AD, and F 'AF. We
already know that EAE = A. To carry out the combination A - 'AA, we first note
from our table that AA = E. Thus, A -'AA = A -'E, so we have to find out what
element is equal to A -'. We know that A -'A = E; accordingly, we go to the A
column and determine which element, when combined with A, produces the result
(matrix element) E. This element is A, and therefore A -1 = A. As a result,
A -'AA = A -'E = AE = A. In a similar fashion, we can show that B '= B. To
determine B-'AB, we see that AB = D and B-'D = BD = C. Proceeding in a
similar fashion, all of the similarity transforms of A can be determined:
E 'AE = A
A- 'AA = A
B- 'AB = C
C- 'AC = B
D- 'AD =C
F-'AF = B
Thus, we have determined that A, B, and C are all conjugate and members of
the same class. We now know that all transforms of B and C are also A, B, or
C (see rule 3, p. 19), so these do not have to be worked out.
Next, we find an element that is not conjugate with A and determine its
transforms. Such an element is D, and its transforms are:
E-'DE=D
A-'DA = F
B-'DB = F
C-'DC = F
D-'DD=D
F-'DF= D
All transforms of D are either F or D, so these two elements constitute a class.

Note that we have now placed every element in a class.
2-4 Summary of the properties of Vectors and Matrices 25
We can again make this procedure less abstract by working out the classes
for the symmetry group of the ammonia molecule from Table 2-1. (You might
try closing the book, working this out yourself, and then using the ensuing
discussion to check your result.)
EEE = E
2
C3 EC3 = E (note: C 3 -1 C 32 )
av"Eac" = E
The first result is, of course, that E is always in a class by itself. Next, it can be
shown that C3 and C 32 form a class of the group.
EC 3E C3
2
C3 C 3 C3 = C3
C3 C3 C32 = C3
2
Cca = C3
2
i,C 3 v,' = C3
2
"C3 av" = C 3
Similarly, ac, cr ', and a," form a class of the group:
EaE = a
C3207,C3 = a-,"
2 =
C3 C3
CI C
In general, equivalent symmetry operations belong to the same class. (Recall 1z ---- \*1
that equivalent elements are ones that can be taken into one another by symmetry
operations of the group.) For example, in the planar ion PtCl42- (D4,), shown
in Fig. 2-2, av and av' can never be taken into a, and a,'. Therefore, uv and o-v'FIUE22Teaan
form one class of the group, while a, and a,' form another. U lnsi h ~ t'
2-4 SUMMARY OF THE PROPERTIES OF VECTORS

AND MATRICES
Vectors
A vector in three-dimensional space has a magnitude and a direction that can
be specified by the lengths of its projections on the three orthogonal axes of a
Cartesian coordinate system. Vector properties may be more than three-
dimensional, so the above statement can be extended to p-dimensional space and

B 9 p orthogonal axes in p-space.
It is often necessary to take the product of vectors. One type of product that
A produces a number (i.e., a scalar) is called the scalar or dot product. This is given
(A) by:
AB= AB cos 0 (2-1)
Here the boldface A and B refer to the two vectors, and the dot refers to
their dot product. On the right-hand side of the equation, A and B refer to the
0 lengths of the A and B vectors, and 0 is the angle between them as shown in Fig.
2-3a. Accordingly, if the angle 0 between vectors is 90 , the dot product is zero
(cos 900 = 0) and the vectors are orthogonal.
It is necessary to reference our vectors to a coordinate system. This is done
for a two-dimensional example (i.e., the xy-plane) in Fig. 2-3b. The angle 0 is
now seen to be e - p, so the dot product becomes
x
(B) A-=AB cos (e - (p) (2-2)
FIGURE 2-3 a. Two vectors,
A and B, separated by an Simple trigonometry now gives us the projections of the A and B vectors on the
angle 0. b. The vectors x and y axes as:
placed in a two-dimensional Ax = A cos p (2-3)
xy-coordinate system.
A, = A sin cp (2-4)
B, = B cos g (2-5)
B, = B sin e (2-6)
Using a trigonometric identity, equation (2-2) can be rewritten as
A- B = AB(cos p cos E + sin p sin r)

= A cos cp Bcose+A sin cpBsinE (2-7)
Substituting equations (2-3) through (2-6) into (2-7) produces
A - B = A B + AB,
Thus, the dot product of two vectors in two-dimensional space is the product of
the components with all cross terms (A.B, and so forth) absent. In p-space, the
result obtained is
A-B = AiB
i=1
where i ranges over the p orthogonal axes in p-space. Accordingly, the scalar
square of a vector is given by
p
A2 , A
i=1
Matrices
A matrix is a rectangular array of numbers or symbols that has the following
general form:
a11 a12 a 13
a21 a 22 a 23
a31 a32 a33
The square brackets indicate that this is a matrix, as opposed to a determinant.

The entire matrix can be abbreviated with a script letter or by the symbol [aij].
The symbol aj refers to the matrix element in the ith row and jth column. When
the number of rows equals the number of columns, the matrix is called a square
matrix. The elements ay of a square matrix for which i = j (i.e., a,,, a 2 2 , a 3 3, etc.)
are called the diagonal elements, and the other elements are called off-diagonal.
When all of the off-diagonal elements of a matrix are zero, the matrix is said to
be diagonalized or to be a diagonal matrix. When each of the diagonal elements
of a square matrix equals 1 and all off-diagonal elements are zero, the matrix is
called a unit matrix. The unit matrix is often abbreviated by the Kronecker delta, 5.
1 0 0 0
0 1 0 0
0 0 1 0
0 0 0 1
Unless i = j, the matrix element ij of a diagonal matrix has a value of zero.

The trace or character of a square matrix, an important property (vide infra), is
simply the sum of the diagonal elements. A one-row matrix can be conveniently
written on a single line. In order to write a one-column matrix on a single line,
it is enclosed in braces, { }.
A vector is conveniently represented by a one-column matrix. In a three-
dimensional orthogonal coordinate system, a vector initiating at the origin of
the coordinate system is completely specified by the x, y, and z coordinates of
the other end. Thus, the matrix {x, y, z is a one-column matrix that represents
the vector. In p-space, a p by 1 column vector is needed. In both instances, the
elements of the matrix give the projections of the vector on the orthogonal
coordinates.
Matrices may be added, subtracted, multiplied, or divided by using the
appropriate rules of matrix algebra. In order to add or subtract two matrices 21
and 0 to give a matrix E, the matrices must all be of the same dimension; i.e.,
they must contain the same number of rows and columns. The elements of the
E matrix are given by:
c=a± by
A matrix can be multiplied by a scalar (a single number). When multiplying by

a scalar, each matrix element is multiplied by this scalar
k[aj] = [kaij]
The ijth matrix element of a product matrix is obtained by multiplying the ith
row of the first matrix by the jth column of the second matrix, i.e., row by
column to produce the product matrix. That is, the matrix elements of E,
the product of W and 0, are given by
n
Cik - Y aijbik
j=1
where n is the number of elements in the ith row and in the jth column. This
matrix multiplication is equivalent to taking the dot product of two vectors.
It should be clear that, in order to multiply a matrix by another matrix, the
two matrices must be conformable; i.e., if we wish to multiply W by 3 to give
E, the number of columns in W must equal the number of rows in 2. If the
dimensionality of matrix W is i by j and that of 3 is j by k, then E will have a
dimensionality of i by k. This can be seen by carrying out the following matrix
multiplication:
a1 1 a 12 ] bl, b12 b1 3 C11 C12 C1 3

a 21 a 22 = C2 1 C2 2 C2 3
a31 a32 b2 1 b2 2 b2 3 C31 C32 C33
3 by 2 2by 3 3 by 3
The number of columns (two) in W equals the number of rows (two) in 2. The
matrix elements of S are obtained by a row by column multiplication; i.e.,
C1 1 1 bl, + a1 2 b 2 l
C12 ab 12 + a 2 b 22
C13 1 1b 3 + al 2b2 3
C2 1 = a 21 bl + a2 2 b 21
C2 2 = a2 lb 1 2 + a2 2 b 2 2
C2 3 =a2 lb 13 + a2 2 b 2 3
C3 1 =a3 lbl, + a3 2 b 2 l
c = a3 b 2 + a3 2 b22
C3 3 ~~ a3 3 + a3 2 b 2 3
Matrix multiplication always obeys the associative law, but is not necessarily
commutative. Conformable matrices in the order WO may not be conformable
in the order 29.
Division of matrices is based on the fact that Wdivided by 3 equals 92-1,
where 0-1 is defined as that matrix such that 232 = 6,j. Thus, the only new
problem associated with division is the finding of an inverse. Only square matrices
can have inverses. The procedure for obtaining the inverse is described in matrix
algebra books, which should be consulted should the need arise.
Conjugate matrices deserve special mention. If two matrices ') and 23 are
conjugate, they are related by a similarity transform just as conjugate elements

of a group are; i.e., there is a matrix 91 such that
W = 9- '391 (2-8a)
One advantage of matrices that will be of significance to us is that they can

be used in describing the transformations of points, vectors, functions, and other
entities in space. The transformation of points will be discussed in the next section,
where we deal with a very important concept in group theory: how symmetry
operations (elements) can be represented by matrices and what advantages are
thus obtained.
We can use matrices and vectors to gain a further appreciation of similarity
transforms. Imagine three distinct fixed points 0, P1, and P2 in three-dimensional
space. Let X be the vector from 0 to Pi, and let Y be the vector from 0 to P 2 -
A set of three-dimensional Cartesian coordinates centered at the point 0 will be
referred to as frame I. Suppose that an operator A, also thought of as being
associated with frame I, transforms the vector X in frame I into the vector Y in
frame I according to the equation
V = AX
Next rotate frame I into a new position, keeping the origin of the coordinates
fixed at 0 and leaving the points P, and P 2 fixed in space. Frame I in its new
position is called frame II. Notice that the vectors X and Y have been unaffected
by this procedure, since the three points 0, P 1, and P2 have remained fixed in
space. However, the projections of X and V on the coordinate axes of frame II
will give numerical values different from those obtained when the vectors were
projected on the coordinate axes of frame I. Thus, in frame II we write the vectors
as X' and Y'. We now want to find an operator A' associated with frame II that
transforms the vector X' in frame II into the vector Y' in frame II according to
Y = A'X'
The operator A' in frame II is said to be similar to the operator A in frame I.

Thus, A' sends X' (expressed in frame II) into Y' (also expressed in frame II),
while A does the same thing in frame I. In order to evaluate the operator A', we
need to know the relationship between the components of X and Y in frame I
and the components of X' and Y' in frame II. This information is given in the
form of a transformation of coordinates expressed by a matrix S. Thus,
X = SX'
V =SY
so that
Y = AX
becomes
SY'= ASX'
Since IS I # 0, we know that S ' exists, and
Y'=S 'ASX'
Thus,
Y' = A'X'
A'= S- 1AS (2-8b)
The similar matrices A' and A are connected by equation (2-8b), which is known
as a similarity transformation. The trace of a matrix is invariantunder a similarity
transformation.
2-5 REPRESENTATIONS; GEOMETRIC

TRANSFORMATIONS
As discussed previously, E, a, i, C, and S,, describe the symmetry of objects.

S(1 Each of these operations can be described by a matrix. Consider the point P in
Fig. 2-4 with x, y, and z coordinates of 1, 1, and 1 corresponding to the
projections of the point on these axes. The identity operation on this point
X corresponds to giving rise to a new set of coordinates that are the same as the
E1 [X'
old ones. The following matrix does this.
FIGURE 2-4 A point in the

Cartesian coordinate system.
00
1 00] rX YI=IY'
0 0 1 Z Z'
Matrix multiplication yields:
X =X'
Y =Y
Z Z'
This unit matrix is said to be a representation of the identity operation.

A reflection in the xy-plane, o.,, changes the sign of the z-coordinate, but
leaves x and y unchanged. The following matrix does this.
1 0 0 X X'1
0 1 0 Y = Y'
0 0 -1 I Z Z'
Matrix multiplication gives X = X', Y = Y', and Z = -Z', which is the

result of reflection in the xy-plane on the xyz-coordinates of the point P. Similarly,
the result of the operation a.z is given by
010 ]= Y
0 0 1 z Z'
and cr, is represented by
1 0 0 X
Y1
0 -1 0 = XY''1
0 0 1 Z Z'
The inversion operation is given by
0 - 1 0 Y = Y'
0 0 -1 Z Z'
y
Call the z-axis the rotation axis; we shall derive the matrix for a clockwise p (x1,y
(as viewed down the positive z-axis) rotation of the point by an angle p. This
rotation is illustrated in Fig. 2-5. The point (x,, y) defines a vector r 1 that r
connects it to the origin, and rotation leads to the vector r 2 . For any rotation
of r1 about the z-axis by (p, the z component is unchanged, leading to:
0 x x"X2
I z z'
FIGURE 2-5 Clockwise

In Fig. 2-6, the rotation of r1 to r2 is shown. The problem is to determine the rotation of the point (x1, y1)
new coordinates for r 2 . The dots show rotation of the y1 coordinate of ri by (p, by the angle qp.
x component
Y ----------------- r,(x, d FIGURE 2-6 Rotation of the

vector r1 and its y-coordinate
through the angle (p.
y component vector a in
(x2, Y2)
producing the vector y'. The vector y' now has both x- and y-components, which
are given by:
y' = y1 sin (p + y1 cos (p
This result is obtained by the following trigonometric arguments: y' still has
a length of y1. Using trigonometry on the triangle defined by <p, y', and the y-
component of y', we see that
y component y component
cOS (p =-
Y Y1
or y component = y1 cos (p. Similarly, since sin (p= x component/y 1, we have

x component = y1 sin (p. The vector y' then has components given by:
y' = x component + y component = y, sin (p + y1 cos (p
Figure 2-7 illustrates the rotation of the x-coordinate of the vector ri,
producing a new vector x' with both x- and y-components. As before, we can
show that the y-component is -x, sin (p and the x-component is xi cos p. This
leads to:
x'= x, cos (p - x, sin (p
x component
r1 original
vector
FIGURE 2-7 Rotation of the

x-coordinate of the vector r1 by the angle
y component
X1
The x 2 - and y 2-components of the rotated vector r2 must equal the sums of the
x- and y-components of x' and y', so:
x2 = X1 cos (p + y 1 sin (p
Y2 = -X 1 sin (p + y1 cos (p
Writing these equations in matrix form, we obtain for the clockwise rotation of
a point in a fixed axis system:
co s (p sin (p 1x [x 2]
-sin (p cos (py y2
For the reverse transformation (i.e., counterclockwise rotation of a point in a
fixed axis system) the transformation matrix is given by*:
cos <p -sin <p

sin <p cos p]
The total transformation matrix for a clockwise proper rotation is:
cos p sin p 0
sin <p cos p 0
0 0 1
For a C2 rotation we substitute (p = -180 into the clockwise proper rotation

matrix. A negative angle corresponds to a clockwise rotation according to the
trigonometric convention.
-1 0 0
0 -1 0
0 01
A C3 rotation matrix can be written by substituting p = 120 in the clockwise
proper rotation matrix.
For an improper rotation around the z-axis, we rotate in the xy-plane and
then reflect through it, leading to the matrix representation
cos p sin p 0
-sin <p cos <p 0
0 0 -1
The multiplication of these matrix representations of the symmetry oper-

ations produces the same result as the product of the symmetry operations, as
they must if they are indeed the correct representations.
a-z x o-z = E
0 01 1 0 0 1 0 01
0 -1 0 0 -1 0 = 0 1 0
0 0 1 0 0 1 0 0 1
All matrices that describe the tranformations of a set of orthogonal

coordinates* by the symmetry operations of a group are orthogonal matrices.
Their inverses can be obtained by transposing rows and columns. For example,
* This is equivalent to a clockwise rotation of the axis system, which would be written as:
Cos - sin
c P' -
sin a Cos Yp Y
*Orthogonal coordinates are those whose dot product is zero.
the inverse of the matrix that corresponds to a 30 rotation about the z-axis
[cos (-30')= -3/2]
3 1 0 1 0
2 2 2 2
1 3 1 3
2 2is 2 20
0 0 1 0 0 1
2-6 IRREDUCIBLE REPRESENTATIONS
The total representation describing the effect of all of the symmetry operations
in the C2 , point group on the point with coordinates x, y, and z is
E C2 og 0%
Notice each of these matrices is block diagonalized; i.e., the total matrix can be
broken up into blocks of smaller matrices with no off-diagonal elements between
the blocks. The fact that it is block diagonalized indicates that the total
representation must consist of a set of so-called one-dimensional representations.
We can see this as follows. If we were to be concerned only with operations on
a point that had only an x-coordinate (i.e., the column matrix
{x, 0, 0}), then only the first row of the total representation would be required
(i.e., 1, -1, 1, - 1). This is an irreducible representation, which in this case is a
set of one-dimensional matrices describing the symmetry properties of the one-
dimensional x-vector in the specified point group. The symbol B1 will be used
to symbolize this irreducible representation. Do not concern yourself with the
meaning of the B and 1 for now, but just think of this as a label. The irreducible
representation for y is:
1 -1 -1 1
which is labeled B 2 ; and that for z is:
which is labeled A 1 .
The total representation-the four 3 x 3 matrices-is a reducible represen-
tation. It is reducible, as discussed above, into three sets of 1 by 1 matrices, each
set a representation by itself. The trace or character of each of the total
representation matrices is the sum of the characters of each of the component
irreducible representations; in order, these are 3, -1, 1, and 1.
We will show shortly that the block diagonalized 3 x 3 matrices for the
effect of symmetry operations on our point resulted because the x-, y-, and z-axes
2-7 Character Tables 35
of our coordinate system were selected as the basis set of vectors for the symmetry
operations, i.e., z for rotation, xy for reflection. We shall subsequently consider
a representation that is not block diagonalized. Z
There is one additional irreducible representation in the C2,, point group.
We can illustrate it by carrying out the operations of the group on a rotation
R. described by the sketch on the right. The identity and C2 rotation do not
change the direction in which the arrow points, but reflection by ax2 and a do. *
None of the irreducible representations we now have, nor any combination of
these irreducible representations, describes the effect of the C2 symmetry
operations on this curved arrow representing rotation around the z-axis. The
result is another irreducible representation:
1 1 -1 -1
which is labeled A 2 .
2-7 CHARACTER TABLES
We now have empirically derived all of the irreducible representations of the C 2 .

point group, but we cannot be sure we have all of them with the procedures
employed. There are rigorous procedures for deriving all of the irreducible
representations of the various point groups, which are covered in many treatments
of group theory. These procedures will not be covered here, for we will be more
concerned with using the irreducible representations and will not have to generate
them; they are readily available. The results of the preceding discussion for C 2 ,
symmetry are summarized by the character table of the C 2 , point group shown
in Table 2-2.
TABLE 2-2. Character Table for the C2, Point Group

E C2 'xz "$
Al +1 +1 +1 +1 z
A2 +1 +1 -1 -1 Rz
B1 +1 -1 +1 -1 x, R,
B2 +1 -1 1 +1 y, Rx
Each entry in the character table is the character of the matrix for that
operation in that representation. For 1 x 1 matrices, the character and the matrix
are the same. There is a row for each irreducible representation.
The labels have the following general meaning:
1. The symbol A indicates a singly degenerate state (i.e., it consists of only
one representation) that is symmetric about the principal axis: i.e., the character
table contains values of + 1 under the column for the principal axis for all A
species.
2. The symbol B indicates a singly degenerate state that is antisymmetric
about the principal axis; i.e., the value - 1 appears under the column for the
principal axis for all B species.
3. The subscripts 1 and 2 indicate symmetry or antisymmetry relative to

a rotation axis other than the principal axis. If there is no second axis, the
subscripts refer to the symmetry about a a. plane (e.g., in the C2 group, subscript
I is symmetric about the xz-plane and 2 is antisymmetric).
All properties of a molecule with C2 symmetry can be expressed in
mathematical terms that will, for a basis set, form a general representation that
consists of one (or a combination of more than one) of these irreducible
representations. As we shall see, appropriate combinations of the irreducible
representations summarize the results of our matrices operating on anything
belonging to the C2 point group. Character tables are available for all of the
point group symmetries, and are listed in Appendix A.
2-8 NON-DIAGONAL REPRESENTATIONS
If, instead of selecting the C2 and a elements to be coincident with the axes of
. our coordinate system, we selected as our basis the C2 axis shown in Fig. 2-8
C2 axis (and the other elements appropriately), then our matrix representation would not
be block diagonalized. For example, the C 2 rotation would not simply change
the sign of the x-coordinate as we saw before, but instead would have to keep
y the sign the same and change the magnitude of the coordinates. Instead of
-0
0[
X 0 1
FIGURE 2-8 A basis set for
geometric transformations on
the new x-coordinate would be dependent on what the y- and z-coordinates of
p that are not coincident with
the Cartesian coordinates. (P the original point were; i.e., there would be off-diagonal matrix elements. We
has positive x, y, z shall not work this all out, for it is not worth the effort. However, the point
coordinates.) should be made that the trace of the matrices for all of the symmetry operations
would be the same for any Cartesian basis set, i.e., 3, - 1, 1, 1. The matrix
representation for the axis selection we have made here can be converted into a
block diagonal matrix by using an appropriate matrix, Q, for the similarity
transform; i.e.,
Q-'C 2 Q, Q 'iQ, etc.

In this case, the similarity transform is a rotation to the new basis set (for the
symmetry operation with the rotation matrix) from the basis set of the original
X, Y, Z axis system. The basis set is rotated, not the points or the molecule. In
practice, when off-diagonal elements result, we realize that a poor selection of a
coordinate system was made. We choose a new orthogonal coordinate system,
each basis vector of which transforms as one of the irreducible representations,
so that a block diagonal representation will result. This will be made clear by
picking a simpler example that we can work through completely. Such a
representation, which is not block diagonalized, can be obtained by considering
the operations of the C 2 c point group on two vectors a and b, using these vectors
as the basis set to describe the representation. The problem is illustrated in Fig.
Z
-a
-b
b 900
xY
\2,
(1 -/ 2\/,0) a (11/\Z 1 / \/Z_, 0)
X
FIGURE 2-9 Geometric transformations in the a,b basis.
2-9. The coordinates were selected as 1/2 to produce normalized (i.e., unit)
vectors:
)2 (+ ) =
The angle between a and b is 90 . We are going to set up our matrices to operate
on the a, b basis set, so these matrices will not work for the x, y, z basis set of
the X, Y, Z-coordinate system. They will tell us what a and b are changed to, in
terms of a and b. Since we wish to illustrate with this example the dependence
of our matrices on the location of our basis set in the coordinate system, we shall
carry out all of the symmetry operations about axes and planes in the X-, Y-,
Z-coordinate system. In terms of the a, b basis, we have
Ea = a 1 0]
Eb = b [0 1
Now, the C 2 operation changes a to -a and b to -- b
C 2 a= -a - 01
C2b= -b 0 1
Reflection in the xz-plane moves a into b and b into a:
xzba=b [0 1]
Furthermore,
ayZa = -b 0 -1]
C,,b = -a -1 0
Note that these reflections have moved a into b and b into a with off-diagonal
elements, and that the trace is zero. Further note that the trace of their
representation matrices is the same as those for the x, y parts of the matrices
generated earlier in the x, y basis for the C 2, point group. The off-diagonal
elements immediately tell us that we made a bad choice of axes for this basis set.
We know from earlier discussion that x and y form the basis of irreducible
representations in C2,; thus, if we locate our vectors along x and y, and use
matrices that work on the x- and y-coordinates of the point defined by the vector
(i.e., an x, y basis set), then diagonal matrices will result. We can do this with a
similarity transform.
For our first example of this procedure, we shall consider the operation c,..
The similarity transform is Q- 'o,2Q, where Q is a clockwise rotation matrix of
the xy-axis system by 45 (this operation makes x and y coincide with a and b).
[cos
sin
qp
p
sin q
cos (
= Q when p =45'
so
1 11 1
Q=- /2and Q =
Q1l(b)Q- YZ1
[,2- ,/0
1
42 /2 I0
-- K72
0A[1 -1/2
1 1=
-1/ 02
Y
Note that the trace of the resulting matrix is the same as that of the starting one.
The trace of a matrix is always invariant under a similarity transform. When we
operate on all of the matrices for the various operations in C 2 . with this similarity
transform, we will have
E C 2 ., UVX, U,2
01 -1 01
10
[1I 1
-1
0 -1
011
0 -1 0 1
corresponding to the B1 (upper left) and B 2 (lower right) representations for x

and y, respectively; i.e., our reducible representation in the a, b basis is decomposed
into B1 and B 2.
Degenerate Representations
For further practice, consider the ammonia molecule, represented in Fig. 2-10
with an x, y, z-coordinate axis basis set. The symmetry operations are E, 2C 3 ,
o'" Y ',
(-1/2, 3/)H
FIGURE 2-10 The ammonia molecule in

a Cartesian coordinate system. (N is
(1,0) above the xy-plane, and the projection of
the N-H bond on the x-axis is taken as 1.)
(-1/2, - 3/2)H
and 3a, Consider the x-, y-, and z-axes all together:
2
E= 01 01 001
0 0 1 2
0 0 1
The C3 rotation matrix is obtained by substituting (p = 2n/3 (= 1200) into the

general proper rotation matrix defined earlier (p. 33).
In order to determine the matrices for a, we must work out the reflection
matrix for each plane. The matrix for a, is quite simple:
0Y=]Z'[Y'
Ol1 Z]
We see that reflection in a, just changes the sign of y. The following matrices
give the new x- and y-coordinates after reflection by a,' and a,":
1 3 1
3 0
2 2 2 2
0*' = /3 1 a, =-
1 0
2 2 2 2
0 0 0 0 1
To employ these matrices, substitute the x, y, z-coordinates of one of the hydrogens

for X, Y; and Z to get the new coordinates x', y', and z' after the operation.*
not change, and x and y are given by
[fJ
2
3]1 +3
* For example, to determine the effect of o.' on (- 1/2,
2
2 2
,/3- /3
4
/3+
/3/2),
Z:
we see that the z-coordinate will
0
2 2 2 4 4.
Note that a,' does not change (- 1/2, -//2) but does interchange the other
two hydrogens. For this symmetry we have off-diagonal elements that
simultaneously change the old X-coordinate into a new one with x- and y-
components.
Next, we shall work out the irreducible representations for the x, y, and
z axes in the C 3, point group. The z-axis, which is the rotation axis, transforms
unchanged and can be thought of as being described by the three 1 x 1 matrices
[1].
E 2C 3 3a,
[1] [1] [1]
A 1200 rotation of a vector on the x-axis produces a vector with both x- and
y-components (Fig. 2-11). The same thing occurs for rotation of y. Substituting
p = - 120 into the rotation matrix and multiplying by {x, y} gives x' and y' as:
x -()
y3)
FIGURE 2-11 A rotation of

= + ) -(x
- ()
the x-axis by 120 about a
z-axis represented by the dot. As we have just shown, the two vectors, x and y, are related and cannot be
transformed independently. The x- and y-vectors are represented in the character
table by the symbol E, used to indicate double degeneracy. The trace of the
matrix (obtained by summing the diagonal elements, and which has been shown
to be independent of the basis selected) is the character reported in the character
table. The character is invariant under any similarity transformation carried out
on the matrix. The significance of this will be made clearer when we consider
more complicated transformations.
In summary, the characters of all non-degenerate point groups (i.e., those
having a C 2 as the highest-fold rotation axis) are 1 x 1 matrices with characters
of + 1 or - 1, which indicate symmetric or antisymmetric behavior of an
appropriate basis function under the symmetry operation. A degenerate point
group contains a character that is the trace of the transformation matrix and
that summarizes how the degenerate basis set transforms together.
Generally speaking, then, we shall end up with an n x n matrix for our
representation (where n is the size of our basis set) for each operation in the point
group; e.g., in C2,,
If each of these is block diagonal, they correspond to irreducible representations.

If not, we want a similarity transform to block diagonalize them.
2-9 More on Character Tables 41
2-9 MORE ON CHARACTER TABLES
We earlier defined a class as a complete set of elements that are conjugate to

each other. We also mentioned that conjugate matrices have identical characters.
(You can confirm this by checking the trace of the a, a,', and a," matrices just
given for the degenerate x and y vectors.) Thus, classes are grouped into one
entry in the character tables; see, for example, 3c, of the C 3, character table.
The different irreducible representations may be thought of as a series of
orthonormal vectors in k-dimensional space, where k is the number of classes in
the point group. Since vectors that form the basis for two different irreducible
representations, y, and XZ,are orthogonal, we have
Y gyi*(R)Zj(R) = 0 when i # j (2-9)

R
The sum is taken over all h symmetry operations (each denoted by R) in the point
group, and g represents the number of elements in the class. The sum of the
squares of the characters of any irreducible representation equals h, the order of
the point group.
Y
R
g[i(R)] 2 = h (2-10)
where R is a sum over all operations, so 3u, is counted as three times a, in this
summation for g = 3.
A character table (Table 2-3) is divided into four main areas, as explained
in the following paragraphs.
TABLE 2-3. Character Table for the C3v Point Group
C3 , E 2C 3 3c,
A, 1 1 1 z x2+y2,z2
A2 1 1 -1 R,
E 2 -1 0 (x, y) R R, (x 2 _ y 2 , xy)(xz, yz)
I II III IV
Area I: The meanings of A, B, E, and the 1 and 2 subscripts have been

discussed. Other symbols are encountered in the character tables; see Appendix
A for examples. The symbols E and T represent doubly and triply degenerate
states, respectively. (F is often employed instead of T). If the molecule has a
center of symmetry (see the C 2 , D2h, D 4 ,, and 0 , character tables in Appendix
A), the subscript g is used to indicate symmetry (+ 1) with respect to this center,
while u indicates antisymmetry (- 1). Prime and double prime marks are employed
to indicate symmetry and antisymmetry, respectively, relative to a ah plane of
symmetry.
Examples of the application of these rules can be obtained by referring to

the character tables in Appendix A. All A species in the D 3h group have values
of + 1 in the column for the principal axis, 2C 3 . The species A, are symmetric
(i.e., have + 1 values) to 3C 2 , and A 2 species are antisymmetric (-1) to these C 2
axes. In D 4 h, all A species are symmetric with respect to the perpendicular C 2
axes. The prime species are symmetric (+) to the horizontal plane a, while the
double prime species are antisymmetric (-) to this plane. The u species are
antisymmetric (-) to the center of symmetry, while the g species are symmetric
(+).
Area II: This has already been discussed, but there are a few additional
points to be made. Some character tables contain imaginary or complex
characters. Whenever complex characters occur, they occur in pairs such that
one is the complex conjugate of the other. While they are mathematically distinct
irreducible representations, in all physical problems no difference is observed and
the results are as if they had been a doubly degenerate irreducible representation.
In the Td point group, two triply degenerate species exist, T, and T2 . For T2 , x,
y, and z form a basis; for T, all three of the rotations about these axes form a basis.
Area III: This area lists the transformation properties of vectors along the
x, y, and z axes and rotations R,, R,, and Rz (represented as curved arrows)
about the x, y, and z axes.
Area IV: This area indicates the transformation properties of the squares
and binary products of the coordinates. Although there are six possible square
and binary products (x2 , y 2 , z2 , xy, xz, and yz), only five are ever indicated because
x2 + y2 + Z2 = r2 and, as a result, one of the linear combinations is redundant
2
(r2 is the square of the radius of the x + y + z sphere). The direct product of
2 2
two vectors is obtained by multiplying the species for each, e.g.,

XY = B1 x B 2 = A 2 . To perform this multiplication, the following procedure is
used:
E(B1) x E(B 2) C 2(B1) x C 2 (B2 ) a,(Bl) x a,(B2) uv'(B 1 ) x a '(B2)
or
giving the result

+1 +1 -1 -1
The result + 1, + 1, - 1, -1 is identical to the irreducible representation A2;

hence, B 1 x B 2 = A 2 . In subsequent chapters we shall have occasion to take the
direct product of irreducible representations, and this procedure will be employed.
Many of the combinations in this area have symmetry properties identical to
those of the d-orbitals (e.g., xy and d, are identical). Thus, these products can
be used to indicate which orbitals will remain degenerate in transition metal ion
complexes of various symmetries. The pair x _ y2 and xy are degenerate in a
2
C 3, molecule as are xz and yz, but xz and yz belong to a different

class than do
x2 _ y2 and xy. The different classes can have different energies in complexes
with different symmetries.
The wave tunctions of a molecule are an example of a basis for a
representation; that is, they can serve as the basis set for the representation
matrices of a group. These wave functions must possess the same transformation
properties under the operations of the group as the irreducible representations.
2-10 More on Representations 43
Thus, each molecular orbital in a molecule will have a symmetry given by one
of the irreducible representations in the point group of the molecule.
2-10 MORE ON REPRESENTATIONS
If, instead of vectors, we choose to obtain a representation using a set of atom

coordinates, the resulting matrix could have dimensions of 3N x 3N associated
with the three Cartesian coordinates for each of the N atoms in the molecule
(Fig. 2-12).
FIGURE 2-12 The 3N coordinates for

H20 in a Cartesian coordinate system.
Zb Ix
Yb
X
H
H Y, H LYa
For the E matrix, we have a 9 x 9 representation
Any 1 x 1 diagonal matrix element in this large matrix corresponds to a

non-degenerate irreducible representation in the final block diagonalized matrix.
Accordingly, we shall need nine irreducible representations to describe this system.
For the C2 symmetry operationgn our nine dimensional basis set, we have
Similar matrices can be constructed for o (u,)and o- (a,').

These larger representations can be broken up, i.e., are reducible into the
sum of many of the smaller irreducible ones given earlier for the C2 , point group.
We could proceed by looking for a similarity transform Q that block diagonalizes
the 9 x 9 E, C, v, and u,' matrices. Then the blocks in the diagonalized matrix
would correspond to the irreducible representations for this molecule and in this
basis set. Although this would not occur in the C2 point group (or any point
group with a lower than threefold axis), matrices that cannot be reduced to a
1 x 1 matrix would correspond to degenerate representations. After diagonali-
FIGURE 2-13 Form of a zation, the matrix would have the form shown in Fig. 2-13, and each block can
matrix after diagonalization. be treated separately as a representation, provided that the reducible represen-
tations of the other symmetry operations are similarly block diagonalized. There
is an easier way to solve this problem, and we shall describe it next.
2-11 SIMPLIFIED PROCEDURES FOR GENERATING AND

FACTORING TOTAL REPRESENTATIONS; THE
DECOMPOSITION FORMULA
The trace of the matrices for E, C 2 , v, and a,' operations for the problem
described above is
E C2 a, Y|
XT=9 -1 +1 3
In actual practice, similarity transforms are not used to find the irreducible
representations that constitute this representation. Even the total representation
is found by simpler procedures, and this is then factored into the individual
irreducible representations by using the decomposition formula, vide infra.
With a little reflection, it can be seen that the following rules summarize
what we have already said about generating the character of the total represen-
tation.
1. Any vector (or part of the basis set) that is unchanged by a symmetry
operation is assigned + 1. (This diagonal matrix element left the coordinates
unchanged in our earlier discussion.)
2. Any vector or part of the basis set that is changed into the opposite
direction is assigned -1. (The diagonal matrix element reflects this: in the C 2 v
example discussed earlier, all changes were either symmetric (+ 1) or unsymmetric
(-1)).
3. A vector or part of the basis set that is moved onto another vector by
a symmetry operation is counted as zero. Recall that in our discussion of the a
and b vectors, the vectors were interchanged in our earlier example of H 2 O with
off-diagonal elements, and the trace of the matrix dealing with Ha and Hb is zero.
These rules give us directly the value of the trace of the matrices we generated
in our earlier discussion of the nine coordinates (3N) of the water molecule.
Applying these rules to these nine coordinates, we see that E gives a trace
for the total representation, Z,, of 9. The C 2 operation moves all Ha coordinates
into Hb and vice versa for a result of zero; while for oxygen, x goes to -x (- 1),
y goes to -y (-1), and z remains unchanged (+ 1). Accordingly, X, for the C 2
operation is - 1. With x perpendicular to the plane of the atoms, reflection in
2-11 The Decomposition Formula 45
the xz-plane moves all H, and Hb vectors for a result of zero. The x and z vectors
remain unchanged, while y goes to -y for a net result of + 1. Reflection in the
yz-plane does not change any of the six y or z vectors (+6), while the three
x vectors go to -x (- 3) for a total (+6 - 3) of +3. Thus, XT = 9, -1, 1, 3
as described before.
When an operation does not interchange atoms but moves them into a new
position that is some combination of the old coordinates (e.g., C 3 on i), the trace
of the matrix must be determined by working out the matrix for the geometric
transformation as was done earlier. For rotations, the rotation matrix can be
employed.
Next, we shall describe how to determine all of the different irreducible
representations that make up the total representation. The following formula
shows how to do this; books on group theory should be consulted by those
interested in the derivation. The formula summarizing how to factor a total
representation, the so-called decomposition formula, is:
a = gy;(R)xT(R) (2-11)
hR
Here a1 is the number of contributions from the ith irreducible representation,

R refers to a particular symmetry operation (or, if there is more than one element
in the class, to the whole class); h, the order of the group, is given by the total
number of symmetry operations in the point group (in determining this, be sure
to count 30-'s as three); g is the number of elements in the class; 1 j(R) is the
character of the irreducible representation; and Xr(R) is the character for the
analogous operation (R) in the total representation. The use of this formula is
best demonstrated by decomposing the total character for the 3N coordinates
of water,
9
XT= -1 1 3
using the C 2 , character table. The value of h is 4, and g = 1 for all symmetry
operations. We sum over the R (in this case, 4) symmetry operations as follows:
1
aA = [gXAi(E)X(E) + gx A(C 2 )xr(C 2 ) + gxZAc)xT() + gXA.(o)XyrkV)]
1
= 4 [1 xlx 9±+lxlx (-1)±+lxlx1 1xl1x 3] =3
1
aAa2 [1 x 1 x 9 + 1 x 1 x (-1)+ 1 x (-1) x 1 + 1 x (-1) x 3] = 1
4
1
aB = [1 x 1 x 9 + 1 x (-1)x (-1)+ 1 x 1 x 1 + 1 < (-1)x 3] = 2
14
1
a, 2 = -[1 x 1 x 9 + 1 x (-1) x (-1)+ 1 x (-1) x 1 + 1 x 1 x 3] 3
4
Thus, we find that the total representation consists of the sum of the irreducible
representations 3A1, 1A2 , 2B 1, and 3B2 . The total representation above describes
all of the degrees of freedom of the molecule. In infrared spectroscopy, we are

concerned with all of the possible vibrations that a molecule can undergo. If one
subtracts the irreducible representations for the three basic translational modes
and three rotational modes from the total representation, the remaining irreduc-
ible representations describe the vibrations.
A physical view of the decomposition formula can be had by considering
the irreducible representations as vectors in a given point group. The general
representation is also taken as a vector. The dot product of the general
representation vector with an irreducible representation vector that does not
contribute will be zero. The group order is used as a normalization factor to
insure that the dot product of a contributing irreducible representation vector
0
will give the number of contributions from the vector.
Another basis set, used to give a more transparent view of certain molecular
vibrations, employs the bonds to be stretched during a vibration as vectors.
Consider the two vectors for stretching the OH bonds of water, illustrated in
H H Fig. 2-14. The matrices using these bond vectors (S, and S 2 ) as the basis set are:
FIGURE 2-14 Vectors
representing the two bonds
[1 0]
in H20. [1
EE=0 0]
1 0 1
1]
[0
C2=1 01]0 1 0
The total representation X7 is given by:
E C2V(XZ) v(yz)
2 2 0
XT= 0
Factoring the total representation for these bond stretches leads to the
symmetry of the two stretching modes as A1 and B 2 , which correspond to the
symmetric and asymmetric stretch, respectively.
2-12 DIRECT PRODUCTS
We have already mentioned that the characters of the representation of a direct

product are obtained by taking the products of the characters of the individual
sets of functions. In C 2 ,,
B1 x B1 = XE(Bl) X XE(Bl) Xc2(B 1 ) x xc2 (Bl) X (B1) x Xy(B1) X,,(B 1) x X ,(Bl)

= 1 1 1 1
= A1
That is, B1 x B 1 = A 1 . When dealing with degenerate irreducible representations,

remember that the degenerate bases are transformed in pairs, with the character
corresponding to the trace of the transformation matrix for both. This can be
Exercises 47
illustrated by the product of E x E in the C4, point group. In terms of the x, y

basis for C,,, we have
=-1 _0]
0] [ 04 -1]
E=[10 1 4 1 0 2 0 -1
1 0 0 0
[1
ExE=B~~=
E 0 1] X0 1] 0
0 1 0 0
0
1 0
1 0
0
0 0 0 1-
By analogy to taking the cross product, the identity gives a total representation
of four because the cross product of two matrices increases the dimensionality
of the matrix. The formula for taking the cross product of two matrices involves
taking the a, element (where a is the left matrix and b is the right) times the b
matrix to give the upper left 2 x 2 part of the resulting 4 x 4 c matrix; that is,
alibi1' = ca, ab = C1, ab = C2, and a1b2 = c 22 . The a 12 element times
the b matrix gives the upper right quarter of the c matrix, the a21 element times
the b matrix gives the lower left quarter, and the a 2 element times the b matrix
gives the lower right quarter. The character of the direct product matrix is the
product of the characters of the original matrices. In general, the direct products
of irreducible representation matrices may be reducible.
M. Tinkham, "Group Theory and Quantum Mechanics," McGraw-Hill, New York, 1964. ADDITIONAL
F. A. Cotton, "Chemical Applications of Group Theory," 3rd ed., Wiley-Interscience,
New York, 1990.
READING
M. Orchin and H. H. Jaff&, J. Chem. Educ., 47, 246, 372, 510 (1970).
M. Orchin and H. H. Jaffe, "Symmetry, Orbitals, and Spectra," Wiley-Interscience, 1971.
C. D. H. Chisholm, "Group Theoretical Techniques in Quantum Chemistry," Academic
A. Vincent, "Molecular Symmetry and Group Theory," Wiley, New York, 1977.
For space groups, see: J. D. Donaldson and S. D. Ross, "Symmetry and Stereochemistry,"
Wiley, 1972.
1. Work out all the classes (of symmetry operations) in a two-dimensional equilateral EXERCISES
triangle. (Only x and y dimensions apply, so there is only a threefold rotation axis
and 3u, planes.)
2. Work out the group multiplication table for NH 3 . Determine all classes.
3. Identify the subgroups in G,', G,2 , and G6 '. (The groups were given in this chapter.)
4. Demonstrate that the E. and B2 g irreducible representations of the D4, point group
are orthogonal.
5. In the C4, point group, indicate the operations that comprise
a. the 2C 4 given in the character table; that is, which of C 1 , C4 2 , C4 ', or C 44 comprise
2C 4 .
b. the 2a..
c. the 2
ad.
d. Construct a group multiplication table for the symmetry operations in C4 .
e. Can a molecule with C4 . symmetry have a dipole moment? If it can, where is it

located?
6. In addition to vectors and coordinates of a point, the orbitals of an atom can be

assigned to irreducible representations. Using the character table for the D4, point
group, indicate the representations for the p-orbitals and d-orbitals. Take the center
of the orbitals as the point about which all operations are carried out. Why do the
p-orbital representations have a u subscript?
7. For which irreducible representation do the following vibrational modes form a basis?
F -- 2- - 2-
I/F ClC C Cl 2
P \ Pt / 2 Pt
FF l Cl Cl Clj
(a) (b) (c)
8. Recall the problem of a point in the Cartesian coordinate system with an x, y, z basis set.
a. Write the matrix for an inversion operation.
b. Perform the matrix multiplication that shows
i X =
-r C2 -
c. Complete the multiplication table for the C2 1 point group.
9. Consider the C2 . point group in the xy plane. Using unit vectors along the x- and
y-axes as the basis, one finds the following representation:
E C2 an un
[10 01-1
1 0 -1
011
0 -1
01-1
0
01
1
where C 2 is along the z-axis.
X X
Exercises 49
Consider the axes x' and y', which are at an angle of a = 30 counterclockwise with
respect to the x- and y-axes and in the same plane. Keeping C2 along the z-axis and
the - planes in the xz- and yz-planes, give the representation for which unit vectors
along x' and y' serve as a basis set. Show that the traces are the same in the new
representation for each operation as those in the representation above.
10. Generate the matrices describing how the sets of functions (x, y, z), (xy, xz, yz), and
22 2 2
(2z X _ y 2 _ y ) are transformed by the various operations of the group 0.
Show how the trace transforms as irreducible matrix representations designated T1 ,
T2 , and E, respectively.
11. Write the matrices describing the effect on a point (x, y, z) of reflections in vertical
planes that lie halfway between the xz- and yz-planes. By matrix methods, determine
what operation results when each of these reflections is followed by reflection in the
xy-plane.
12. Consider AuCl 4 with D 4 h symmetry. Use the x, y, and z Cartesian coordinate vectors
at each atom as a basis set to form a total representation. Determine the irreducible
representations.
13. In the groups specified, determine the direct product representations and reduce them
when possible.
a. A2 x B1 in C2,
b. E x E in C4 ,.
c. E x E in C,,.
3 3
14. Consider the two complex ions cis-[CoF 4 Cl2] - and trans-[CoF4 C2] -. For each
of these ions, find the symmetry designations of the d orbitals.
15. In problem 10, we worked to find out that in the 0 point group the sets of functions
2 2 2 2 2
(x, y, z), (xy, yz, xy), and (x _ y , 2z _ X _ y ) transformed as T1, T2 , and E,
respectively. The character table contains this information, but is truncated after giving
information on product functions (xy, xz, etc.) and does not show how the triple
product functions (e.g., the xyz) transform. The triple products correspond to the
f-orbital functions (x 3, y3 , z3, xyz, x(y 2 - z 2), y(z 2 _ X2 ), z(x 2 _ y2)). Take all of these
combinations to ascertain how the seven f-orbitals transform in the C2 , point group.
16. Use this square in the following problems. You are to assume that the square is like
a table; i.e., it cannot be turned upside down and still be the same. The two coordinate
systems share a common z-axis perpendicular to the plane of the figure.
Y
450 o'
x
Consider a reflection, ax. (Since the z-coordinate is always constant, work the
following problems as two-dimensional cases.)
a. Derive the matrix corresponding to this reflection in the x, y-axis system.
b. Derive the matrix corresponding to this reflection in the x', y'-axis system.
c. Show that a,' in the new axis system is related to o by a similarity transformation
of the type
a_' = T -xT
Derive the T matrix and verify this equation. The bold symbols represent matrices.
17. a. We discussed the problem of using the OH bonds in water to form a basis for 2 x 2
representation matrices in C 2. symmetry. Consider the H2 S molecule and assume
a 90 bond angle. Write the four matrices for the symmetry operations in C2 , using
the S-H bonds as a basis set.
H 0 H
b. Actually, the same information is obtained from a set of axes 2 and y set up at the
S nucleus as shown, with 2 along the C 2-axis and y in the molecular plane. Write
out the matrices using this basis of I and y.
c. Suppose that R is the matrix for some operation in the z, y frame and that R' is the
corresponding matrix in the S-H bond frame. Then a 1350 counterclockwise
rotation of the points in the plane will convert R to R' by a similarity transformation,
i.e.,
R' = n-'RiT
Write out the matrix nr.
18. The following is excerpted from Strommen and Lippincott J. Chem. Ed., 49, 341 (1972):
In a recent issue it was pointed out by Schafer and Cyvin that the well-known group
theoretical equation
n(,y) YZXR1
9 R
fails to work when applied to linear groups. Equation (1) is used to determine the
number of times, n(y), that a given symmetry species y will occur in the reducible
representation of a particular molecule. XR is the character of the reducible represen-
tation F; X,(') is the character of the irreducible representation; and R is the index
used to denote each of the symmetry operations of the group.
Outline of Method
1. Assume a lower molecule symmetry which corresponds to a subgroup G of the
molecular group G'.
2. Place a set of cartesian coordinate vectors on each atom. (The z-axis must be placed
along the maximum symmetry axis of the parent group.)
3. Using standard methods obtain the characters of F, dcbibe.
Exercises 51
4. Calculate the values of each n(y) by application of equation (1).
5. Finally compare the basis vectors of the irreducible representations under Go with
those obtained for the molecule assuming G.
(By "basis vectors," rule 5 means the functions x, y, z, x2 2

_ y , etc., found in the
right-hand columns of the character table.)
a. To what symmetry group does the molecule acetylene (C2 H2 ) belong?
b. Why does equation (1) fail to work for groups to which linear molecules belong?
c. Use the method outlined (with G = D2 ) to obtain the irreducible representations

(in the group to which C2 H 2 belongs) for a basis set consisting of x, y, and
z vectors located at each atom.
Molecular Orbital
Theory and Its
Symmetry Aspects
Introduction 3-1 OPERATORS
As is true for many of the topics in this text, whole books have been written on
the subject of this chapter. This chapter will be a very brief presentation of those
fundamentals of molecular orbital theory that should be a part of the background
of any modern chemist. We shall concentrate on those aspects that will be needed
for our considerations of spectroscopy. One can well appreciate that a complete
understanding of the spectroscopy, reactivity, and physical properties of substan-
ces will not be possible until we can interpret these phenomena in terms of the
electron distribution in the molecule.
In principle, all the information about the properties of a system of N particles
is contained in a wave function, 0, which is a function of only the coordinates
of the N particles and time. If time is included explicitly, 0 is called a time
dependent wave function; if not, the system is said to be in a stationary state.
The quantity O** (where 0* is the complex conjugate of 0) is proportional to
the probability of finding the electron at a particular point.
For every observable property of a system, there exists a linear Hermitian
operator, &,and the observable can be inferred from the mathematical properties
of its associated operator. A Hermitian operation is defined by
dT
f0*i#j dT = f 0(ifi)*
We shall abbreviate the above integrals by the so-called bra < I and ket > notation
as:
To construct an operator, &,for a given observable, one first writes the classical
expression for the observable of interest in terms of coordinates, q, momenta and
time. Then the time and coordinates are left as they are and for Cartesian
coordinates the momenta, Pq, are replaced by the differential operator, i.e.,
Pq = - ih(0/8q)
where q is the coordinate that is conjugate to Pq. For example, the quantum
52 mechanical operator for kinetic energy, ', can be constructed by first writing the
3-1 Operators 53
classical expression for the kinetic energy:
11 1
T =-mv2
2 +-mv,
2 + my,
2
Momenta = mv = p, so,
1
T- (p, 2 + p 2
+p 2
)
2m
T-
1 -- h
) 2ih + 2
+ -_ih--
)2]
m ax ay ± Z
h2 02 02 02 h2
=
2m
+ aJ]y +
i,z
)-
2m
2
where
02 02
V2=82
ax2 2
ay z2
If an operator &corresponds to an observable, and if 0r, the wave function
for the state, is an eigenfunction of the operator &,then &0, = a. /, where as is a
number. An experimentalist making a series of measurements of the quantity
corresponding to &will always get as. The wave functions are eigenfunctions of
the Hamiltonian operator which yields energies as eigenvalues, i.e., numbers;
ffo = Egn (3-1)
The operator for angular momentum about the Z-axis is also such an operator:
Lifp = mho. (3-2)
where m, the azimuthal quantum number, has values of +1 ... 0 ... -1.
Some properties of a system are not characterized by an eigenfunction for
the appropriate operator which describes that property. Then a series of
measurements of this property will not give the same result, but instead a
distribution of results. The average value or expectation value, <a,>, is given by:
(a,)> =0 1 (3-3)
i is generally not an eigenfunction of the operator &,but using equation (3-3),

one can obtain the average value.
In the course of this book, we shall have occasion to consider many operators.
We shall discuss them as the need arises. Our immediate concern will be with
the Hamiltonian operator and the Schrbdinger equation:
ho = E0
This equation cannot be solved directly for the energy, E, on molecular systems
because one cannot find a function 4 such that fR/O is a constant and not a
54 Chapter 3 Molecular Orbital Theory and Its Symmetry Aspects
variable function of the position of the electron. Consequently, we average fto

over all space by multiplying by i* and integrating over all space:
i* f dr = f *E0 dT Efr**f d
so
E f _ (3-4)
0*0 dr (0 I )
The Hamiltonian operator for most systems is easily written as
H = -(h 2 /87i2 m)V2 + V,
where V is the potential energy of the system. Other operators are just as easily
written by writing the classical expression for the property of interest and replacing
the momentum by -iha/0. Thus, if we had accurate wave functions and could
solve the resulting integrals, we could calculate all of the properties of a molecule.
Unfortunately, such calculations are rarely practical.
In the LCAO-MO method (linear combination of atomic orbitals-molecular
obital) of solving this problem, the wave functions, Oj, of the resultingj molecular
orbitals are approximated as linear combinations of atomic orbitals:
0- = j, (3-5)
where Cjr is the coefficient describing the contribution of p, to the jth molecular
orbital. The total electronic ground state wave function T is a product of the
individual molecular orbital wave functions. As we shall show shortly, the 0''s
form the basis for irreducible representations of the molecular point group.
Therefore, the symmetry of the total electronic ground state wave function is
obtained by taking the direct product of the irreducible representations of the
occupied Oj's.
The various ways of carrying out molecular orbital calculations correspond
to different approximations that are used in the solution of the problem. One
such approximation ignores the correlation of the motion of the electrons in the
molecule. If there are two electrons in a molecule, we know that at any instant
one of the electrons will tend to avoid the region occupied by the other. Over a
longer time period, the motion of one electron will be correlated with the motion
of the second electron. When this correlation is ignored, the resulting Oj's describe
the system in terms of the coordinates of a single electron. Such wave functions
are labeled one electron molecular orbitals. In describing the molecule, two
electrons are added to each of the one electron m.o.'s.
For clarity, we can proceed with the problem of determining our wave
functions by using a two-term basis set Pi and 92 (i.e., two atomic orbitals on
two different atoms) without any loss of generality, i.e., equation (3-5) for our
wave function is
= C 19 1
- + C 2 92
3-1 Operators 55
Substituting this equation for 0 into equation (3-4) produces
j(C1pP1 + C 2 (p2)H(C1<p 1 + C 2 p 2 ) dx
j(C 1 + C 2 (p2)(C1p 1
1 <p
+ C 2qp2) dx
f(C 1<p iC
1 HC 1 p1 + C1<p
1 2 AC1p 1 + C2 (p2 ftC2 tp2j
2 (p2 + C2 <p
dr
fjC 1<1C1p 1 + C2 p2 C1p 1 + C1<1C2 p2 + C2 (p2 C2 p 2] dc
Since C1 and C 2 are constants, they can be taken out of the integrals. We can
abbreviate j9ipfip dr as Hij and fjpip dx as Sij. Furthermore, H 21 = H 12 and
S 12 = S 2 1 , leading to equation (3-6).
E C122H11 + 2C 1 C 2 H12 + C222 H 2 2 (3-6)

C 1 SJJ + 2C 1 C 2 SI 2 + C 2 S 22
The various types of integrals are given different names:
Hi1 integrals are referred to as Coulomb integrals;

H as resonance integrals; and
S as overlap integrals. The value of Sij = 1 for normalized a.o.'s.
The variationalprinciple is employed to find the set of coefficients C, and C 2

that minimizes the energy, i.e., gives the most stable system. The energy we get,
called the variation energy, will depend upon how good our wave function is.
Evar > Eo(ground state energy) (3-7)
This minimization (i.e., the variation procedure) is carried out by setting the
partial derivative of E with respect to each of the coefficients (C's) equal to zero.
Taking the partial derivative of equation (3-6) with respect to each of the
coefficients, setting BE/aC 1 and OE/aC 2 both equal to zero, and collecting terms,
we get
C 1 (Hjj- ES1 1) + C 2 (H1 2 - ES 12 ) = 0

C1H21 - ES 2 1) + C2 (H2 2 - ES 22 ) = 0 (3-8)
This gives us two homogeneous linear equations in two unknowns. The trivial
solution is C 1 = C 2 = 0, but this would require # = 0 (where 0 = Citpi + C2 p 2 )-
Other solutions exist only for certain values of E, and these values of E can be
found by solving the secular determinant:
H11 - ES 11 H 1 2 - ES12 - (3-9)

H 21 - ES 2 1 H 22 - ES 22
The determinant is solved by standard mathematical procedures. Essentially, we

are solving two linear, homogeneous equations in two unknowns. An n x n
determinant gives an nth-order polynomial equation in energy with n roots. Since
the size of the determinant, n, equals the number of atomic orbitals, we get n
energies or n molecular orbitals from n atomic orbitals. In the case where the
two atoms are the same, we get
(H,1 - E)2 - (H1 2 - S 12 E) 2 = 0
(with S, and S22 = 1). The two roots for this equation are:
Ei = and E2 - H11 H 12 (3-10)

1 + S12 2 _S12
Solving or approximating the integrals enables one to solve for the energies.
The energies so obtained can be substituted one at a time back into equation
(3-8), and the coefficients are then obtained by using a third equation:
C1 2 + C 2 2 = 1. The two equations that result are not independent (the deter-
minant equals 0), so we need the third equation.
The evaluation of the S1 2 integrals is straightforward when the structure of
the molecule is known. The integrals H, and H 1 2 cannot be determined
accurately for complex molecules.
For the hydrogen molecule, the Hamiltonian operator is
Y
1,2
(-h 2
/8r2 m)V, 2 - e2/rA1 - e2 /rB - e 2 e 2
B-2 AB + 12
The two hydrogen nuclei are labeled A and B and the two electrons 1 and 2.
The charge on the electron is e, the distance from nucleus A to electron 1 is rA 1'
and so forth. We solve the integrals by using the various individual terms in the
Hamiltonian separately. Accordingly, for one jpftp,
dT, there are many integrals
to be evaluated. The electron-electron repulsion integrals <pi Ie 2 /r 2 I (pj> are
particularly difficult. You can well imagine how complicated the Hamiltonian is
and how many integrals would be needed even for PCI. Ab initio calculations
do all the integrals and thus are practical only for small molecules or wealthy
chemists.
The various ways to "guesstimate" values for these integrals correspond to
various ways to do molecular orbital calculations. We shall return to a more
detailed description of some of these procedures subsequently.
3-2 A MATRIX FORMULATION OF MOLECULAR

ORBITAL CALCULATIONS
Modern approaches to the solution of eigenvalue problems employ a matrix

formulation. Since the quantum mechanical description of many of the spectro-
scopic methods treated in this text is discussed in matrix terms, the previous
section will be reformulated in matrix terms to introduce the general approach.
The solutions of the secular equations are eigenvalues of the Hamiltonian operator
R, so kql = E0 can be written in the LCAO approximation in matrix form as
[fti][C] = [S] [C] [E] (3-11)

3-3 Perturbation Theory 57
where [C] is the coefficient matrix, [Hi] is the energy matrix of elements
(pi p [S] is the overlap matrix, and [E] is the diagonal matrix of orbital
I>,
energies, one for each m.o. The [E] and [C] matrices are to be determined.
Dividing both sides by [C], we obtain
[C] -'[R][C] = [S] [E] (3-12)
The coefficient matrix is, in effect, a similarity transform that diagonalizes the
[Ni;] matrix to produce a unit [S] matrix. The energies result directly from the
diagonalized [h] matrix, as do the LCAO coefficients of the corresponding
molecular orbital.
3-3 PERTURBATION THEORY
Perturbation theory is a useful technique in quantum mechanics when one is

trying to ascertain the effect of a small perturbation on a system that has been
solved in the absence of this perturbation. Basically, we have the solution to
Hf = Ei
We wish to determine the effect of some perturbation on this ith molecular orbital,
i.e.,
R = Eji (3-13)
where
Rt=fl + R' (3-14)
and the term R' is a small perturbation on the system. It can be shown that the
first-order correction to the energy is given by:
Ej1 =<0j' 'f > (3-15)
The correction to the old wave function, 0',can be written as a linear combination
of the k unperturbed wave functions of the original system:
0i =I a iI4 (3-16)
By mixing the appropriate empty excited-state wave function into the ground
state, it is possible to polarize or distort the ground state molecule in any way
we desire. The contributions that the various ,"k functions make in our perturbed
system are given by the coefficients aik, where
- EOkIRI Oi>
0 (k#i) (3-17)
Eko - Ej
Thus, our final expressions for the energy and wave function of the perturbed
orbital, E, and 0,, correct to first order, are
E = Eo + (<j'Ift'l > (3-18)
and
k= +' E OEo> k (3-19)

k oi Ei -E
Note that the zero-order wave function gives us the first-order energy.
The formula for the second-order correction to the energy is given by
0 (3-20)
E | k
k~iEi Ek
It is informative to examine what we have done in the context of a rigorous

solution of a 2 x 2 secular determinant. A 2 x 2 problem is selected because its
secular determinant can be solved easily. We shall again use our original wave
functions and the new Hamiltonian R. We write <'1 |$li0 > as H, and
< 2 I|$' 2 > as H 2 2 , leading to
H1 -E H 12 0
H 21 H22 E
which in turn leads to
E2 - (H 11 + H 22)E + (H 11 H 2 2 - H1 2 2) = 0
with roots
2
/H,,2 + H 2 2
2
- 2H 1 H 22 + 4H 1 2
H1 1 + H 22 ±
2
Ignoring H 1 2 , we obtain the solutions
El = H11 = Ei + <011 |'l 1 > and E2 = H 2 2 = E2 + 0 2' It'l 0 2 >
Using the perturbation equations given above, we write directly
E1 = E1 + (#10 Ift'l 01 > and E2 = E20 + (0 2 If 2 0>
We now see that first-order perturbation theory is permissible only when

the off-diagonal elements can be ignored, compared with the diagonal ones.
3-4 Wave Functions as a Basis for Irreducible Representations 59
The second-order correction to the energy gives us:
H 2
El = H I-H H11
22 -
H 2
E 2 =H - H
22 H22
which is an even closer approximation to the exact solution.
3-4 WAVE FUNCTIONS AS A BASIS FOR Symmetry in

IRREDUCIBLE REPRESENTATIONS Quantum
Since the energy of a molecule is time invariant and independent of the position Mechanics
of the molecule, the Hamiltonian must be unchanged by carrying out a symmetry
operation on the system. (Recall that the symmetry operation produces an
equivalent configuration that is physically indistinguishable from the original.)
Thus, any symmetry operation must commute with the Hamiltonian; i.e.,
HR = RH. Like the irreducible representations, the eigenfunctions are constructed
to be orthonormal:
f OJ*qf dz = o
i.e., 6 = 0 when i # j and 6 = 1 when i = j. It is a simple matter to show that the
wave functions form a basisfor irreducible representations.
If we take the wave equation:
and multiply each side by the symmetry operation k, we get (for non-degenerate
eigenvalues):
ftki =iko
where R~I = fN!, and E commutes with k because E, is a constant.
Consequently, N#, is an eigenfunction. Since Oi is normalized, R#i must also be,
for we do not change the length of a vector by a symmetry operation; therefore,
kfi = ±11i
We have just shown that by applying each operation of the point group to O, (non-
degenerate), we generate a representation of the group with all characters equal
to + 1, that is, a one-dimensional and therefore irreducible representation. Thus,
the eigenfunctions (non-degenerate) for a molecule are a basis for irreducible
representations. It can also be shown that this is true for degenerate eigenfunctions.
Any integrals involving products of molecular orbitals that belong to two

different irreducible representations must be zero. To show this, let fr and f, be
such molecular orbitals that are eigenfunctions of R with eigenvalues a and b.
Thus, (0a I " 1 0 = b(. Ifl 10'>. Since ftA = AH, we have
(0- 1HAk|1 - <0| A|)

100- - a| I fb)
|
A (like other operators with physical significance) is Hermitian, so
(K4I Ht 10||@ - a a l |0
We have just shown that
a R |Ob> = b(oJa I |IO)b>; since a # b, the integral <, f)I must be zero.
3-5 PROJECTING MOLECULAR ORBITALS
In quantum mechanical calculations, we begin with LCAO's that themselves

form a basis for an irreducible representation. Although this is not necessary, it
is clearly often convenient to do so. We shall show next how symmetry
considerations alone provide much information about the atomic orbitals that
contribute to the various molecular orbitals in a molecule. In the last chapter,
it was shown that any mathematical function (e.g., points, 3N Cartesian vectors,
or bond vectors) could be used as a basis set for a reducible representation. It
is a simple matter to extrapolate to the case in which the valence atomic orbitals
for a given molecule are used as a basis set for a reducible representation. As a
simple example, we shall work out the traces of the reducible representation for
the three a.o.'s comprising the ir-orbitals of the nitrite ion, shown in a coordinate
system in Fig. 3-1.
Recall that earlier, when we carried out symmetry operations on the two
vectors a and b, many symmetry operations transformed one vector into the
other and the trace of the representationmatrix was thus zero, the only non-zero
elements being off-diagonal. Accordingly, referring to Fig. 3-1, 0 is moved into
03 with C2 and a, so contributions to the total character from the oxygen
orbitals are zero. This is true whenever atoms are interchanged. The C2 and Uoz
operations on the nitrogen p2 orbital lead to - 1 and + 1, respectively. Reflection
in the yz-plane changes the sign of the p orbitals, making a contribution of -1
for each of the three p orbitals. The total representation is (C2 . point group;
recall that x is perpendicular to the plane).
C2 , E C2 u,(xz) u,'(yz)
3 -1 +1 -3
Factoring the total representation gives A 2 and 2B,; i.e., the symmetries of the
three m.o.'s (from three a.o.'s, we must get three m.o.'s) that result from these
a.o.'s must be A 2 , B1, and B1.
We can now use a technique involving projection operators, P, to give us
more information about the wave functions. The projection operators work on
the basis set (in this example, the three p orbitals) and annihilate (convert to zero)
any element in the basis set that does not contribute to a given irreducible
representation. In using the projection operators, we generally work on one
atomic orbital (or a linear combination of them) (e.g., <p, the p orbital of atom
1 of Fig. 3-1) at a time with the operator
P =hx RR (3-21)
R
where k is the symmetry operation, XR is the4xcharacter of that operation in the

appropriate point group, 1is the dimension of the irreducible representation, and
h is the order of the group. The sum is taken over all symmetry operations in Y
the point group. Accordingly, if we operate on (p, of the nitrite ion with the FIGURE 3-1 A.0. basis set
projection operator of the A 2 irreducible representation, we get: for the 7r-orbitals of the nitrite
ion. (x is perpendicular to
(for A 2 ) = P(A2 )(P - [(1)Ep, + (1)C2 (Pi + (- 1)o(P1 + (- 1) 1 hN'pP1
The numbers in parentheses are characters of the A 2 irreducible representation.

Next, we carry out the specified operations on 9,, and indicate the result in terms
of what the orbital is changed into. The result is fp,
= N[(p, - P31
The result is the molecular orbital corresponding to the A2 irreducible represen-

tation. We shall ignore the factor '/,,, for we will want to obtain the absolute
values for the normalized
4 wave function, and not the relative values resulting
from the operators.
If we had selected the orbital n2 to operate on, we would have obtained
P(A2)(P = 2 + (1)(- 0)P2 + (-

[()(l9P 1)(1)P2 + ( 1)(- 0921 0
This result could mean that there is no A 2 , but we know better, from having
factored the trace of the reducible representation. In this case, a node exists at
atom 2 in the m.o. with A 2 symmetry.
In order to normalize the wave function that we have just obtained, we
realize that
f (A2)9P C= 2
()
dr = C1
2
f 2 d P-
2
2C1) dT C1 2 )932 d 1
3 have the same coefficient. Using an orthonormal basis

Recall that p, and mP set.
d = C12
(p - CP3)2 d 3 2 dT
P123 d -- 2C12 and(Pa = 1.
These substitutions convert the normalization equation to
2C 12 = 1 or C 1
-
/2
The normalized function is
1
(9P1 - (P3)
/2
From factoring the total representation, we know that there is no A 1 symmetry

m.o. We shall use P(A1 ) to show that a consistent result is obtained. First, consider:
P(A )(p i= - [(1)1)9p + (1)(- 1)cP3 + (1)(1)(P3 + (1)(- 1)(P1] = 0

4
Try the other p orbitals, and you will see that there is no A1 symmetry m.o. Next
try P(B2) on (pi:
P(B2 )(P
1 = - [()PI + (+ 1)93 + (- 1)(P3 + (-1)(p1] = 0
4
We know that there are two m.o.'s with B 1 symmetry. Accordingly,
P(B1)cp
i (1)(1)(p, + (- 1)(-- 1)(P3 + (1)(1)(P 3 + (-- 1)(-- 01)(1]
4
SN[(<p + 931
When we use P(B1 ) on P2 the following result is obtained:
P(B1 )92 =- [(1)(1)P2 + (- 1)( 1)92 + (1)(1)92 + (- 1)(- 1)921 = P2

4
Thus, the projection operators produce two m.o.'s of B, symmetry,

- N[pi + 93] and f = 92. The A 2 and two B 1 molecular orbital wave
functions that we have generated are referred to as symmetry adapted linear
combinations of atomic orbitals. Since the B1 m.o.'s have similar symmetries, they
can mix so that any linear combination of these two m.o.'s also has the appropriate
symmetry to be one of the two B1 m.o.'s; i.e., the two m.o.'s can be:
01 = ap 1 + b<P 2 + a(p3
02 = a'<p, - b'<P 2 + a'<p3
The magnitudes of a and b will depend on the energies of the p1 and 92 atomic
orbitals and their overlap integrals. Thus, we cannot go any further in determining
the wave function without a molecular orbital calculation. We shall indicate
approximate ways of doing this in the next sections. We can at this point make
one further generalization. Since there is no symmetry operation that interchanges
FIGURE 3-2 Shapes of the

it-molecular orbitals of the
nitrite ion.
the nitrogen orbital with oxygen, the nitrogen must form a separate irreducible
representation.
We are now in a position to draw pictures of the m.o.'s we have constructed.
They are shown in Fig. 3-2.
If one carries out the operations of the C2 , point group on the m.o.'s shown
in Fig. 3-2, it will be found that these shapes transform according to the symmetry
labels.
We shall work through the C 3,, ammonia molecule as our next example, for
it has more than one element in some classes and also has a doubly degenerate
irreducible representation. The coordinate system is given in Fig. 3-3. Our
procedure is as follows:
1. First work out the total representation and factor it:
Orbital Shapes E 2C 3 3a7, H
7
z
FIGURE 3-3 The location of

NH3 in a coordinate system.
The xz-plane passes through
3 0* 1
y H5 and N.
x
3 0 1
Total 7 1 3
13
*C3 xi I
2
+
2
//3
0 X
3 1
2 2
zj 0 0 1 L
(trace is 0)
64 Chapter 3 Molecular Orbital theory and Its Symmetry Aspects
Factoring the total representation, we obtain:
1
aA =--[1 x 1 x 7+2 x 1 x 1+ 3 x 1 x 3]= 3
6
1
aA 2 =-[lxlx 7+2 xlx1+ 3 x (1) x3]=0
6
1 2
aE x 2 x 7+ 2 x (-1) x 1 +0] =
6
No symmetry operation can take a hydrogen atom into a nitrogen atom, so we

can factor the total reducible representation formed from the hydrogen orbitals
(3, 0, 1) and find an A, and E. In this way we can work on the nitrogen atom
and hydrogen atoms separately.
2. Next, work with projection operators PA and PE. There is no need to
work with PA , for factoring the total representation told us that there are no
molecular orbitals with this symmetry. For this purpose, we label our atomic
orbitals as follows:
(p, = N2s
<P7-= N2p,
(p3 = N2p,
4 = N2p,
(P
(s = H 5 ls (a. contains H,)
(P6 = H 6 ls (a.' contains H,)
(p7 = Hyls (a," contains H 7)
Pon 9s
1
/, - [1Ecp, + 1C 3 9p5 + 1C 3 2(P5 + lae(p5 + 1O,'(P5 + 1"P5]
6
= [(p5 + (p7 + (P + p5+ (p6+7] - [295+ 296 + 2p7]
PA, operating on (p, and p7 will, of course, give the same wave function, so there
is no need to carry these out. We need linear combinations that are not identical;
i.e., they must not be related to one another by multiplication by -1 or by
interchange of the labels of equivalent atoms.
PA, on (p4. None of the operations take p4 into anything else, so it is not
necessary to use the projection operator on p4. It is a basis set by itself and must
3-5 Projecting Molecular Orbitals
belong to the irreducible representation A1 , for it is not changed by any of the

symmetry operations. If we use the projection operator, we obtain:
1
f2 = [,(4 + <P4 + 94 + 941
-6
4
02 = 4 9ps
on q1. We recognize by inspection that cp, is also a basis set by itself

and must belong to the irreducible representation A,.
4
03 = 4 (pi
f,on92. We know by inspection that cp2 and (P3 form a doubly degenerate
basis set that must belong to an E irreducible representation. Thus, it is not
necessary to carry out this operation. We shall do so here to show that we get
the right answer with the operator and to demonstrate how the symmetry
operations work on a doubly degenerate basis set.
1
6 [1Ecp2 + 1Cc3 p2 + 1C 32(P2 + loo(P2 + laA'p2 + lei"(p
2 ]
6
92 (P39
1 3 1 3 1 3
1P
292 293 2 2 2 2 2 2
For C 3 , we use
cos p sin cp y1y

P2
For a clockwise C 3 rotation, p = 1200; for C 3 2 , we have p = 2400. 6 x
60*
x' = x cos cp + Y sin p H'
y' = -- X sin p + ' cos 9p
For the mirror planes, we use the matrices developed in Chapter 2 and the FIGURE 3-4 Effect of
coordinate system shown in Fig. 3-4. reflections through a. on P2
Likewise, the result of fA, on 93 is 0. and the ammonia hydrogen.
Now we apply the projection operator for the degenerate irieducible repre-
sentation E.
E oP 2 4 = [2Eqp2 - 1C3p 2 - 1C 3 2 p 2
1 1(PP P3(P 1 P P
S{292 -292 2 3] 2 2 ]
PE on (P3 = 5 = (3
1
PE on 9s= [ 2 Ep5 1C[2pP - 1C39 5]
1
= [2(P5- (6 - 97]
3
1J
PE on 9P6 = 1 [2 - 97 - (P]
1
PE on 9 = 3 [297 1
- <p - 963
Our resulting orbitals from p5, 96, and 9 are not orthogonal, and we know that
there are only two wave functions in E. Appropriate linear combinations must be
taken to get an orthonormal basis. There are various ways to do this. One that
works for this system involves working with the matrix elements, instead of with
the trace as we did above. The diagonal elements are:
E C3 C3 2 a",
1 0 0 0 1 0 0 0
2 2 2 2
1 1 1 1
0 1 0 __ 0 _0 -1 0 - 0 -
2- 2_ .- 2_ 2.
where a,, corresponds to PE. and a 2 2 to PEb"
PEa on p, gives:
O= NLE(P5~
E1 2 C3 9P
Cp 5 - 2I C
2
Csp P
3 9 5 ± 16a<'2 1,9
"" 2v(
2 "
= N[2p 5 - 96 - 9,1
PEb on 95 gives:
#b = N[E~ps5 1 2 12
C3(p5 2 C3 (P5--1a,<ps+ - U'(Ps+ 1 -a"P5 =1,
=0
There is a node in Vb at 9,. The result of similar operations with PEb on 96 leads
to the other wave function:
Ob = N(9 6 -- 9,)
One additional way to get wave functions for degenerate irreducible repre-
sentations involves working with the character table for the rotation group involving
only the principal axis, i.e., C..
For ammonia, we would use C3 . In C3, the E representations are given terms
of the imaginaries; e.g., for C3,
2
E C3 C3
where k = exp (27ri/3). Thus,
2
Pa9 6 = iE 9 6 + eC3 P6 + e*C 3 P6 - 96 + W97 + E*9 --
2
Pab6 = 1E9 6 + E*C 3 9 6 + 3 P6 - 96 + E97 + W5 = Ob
If we add the two together, we get one wave function; when we subtract the two
and divide by i, we get the second wave function. Upon addition, we get
296 + (e* + 0)91 + (C+ 8*) 9 5
where e = exp (27ri/h) and h is the order of the rotation group. Abbreviating these
as r = exp (i0) and r* = exp (- i'F) and using the trigonometric identities
exp (i0) = cos (D+ i sin (D

exp (- i'F) = cos (D - i sin 'D
we see that E + E* = 2 cos 'D.

In the C 3 group, the rotation axis is 1200, so 2cos (-120') = 2(- '/2) = -
Accordingly, we have
~Ia-296 - 97 - 9
(This is equivalent to our previous result, fr = 295 - 96- 97, obtained by using
the C 3 , point group.)
To get the other degenerate wave function, we subtract 0Ia' from Ob' and divide
by i:
O' r- 0,' (gE e) __
0+ . 97+ (.
Substituting the trigonometric functions for r and * produces
(r - e*) 2i (E* - e) 2i
. =- sin and .i - s
Since sin (- 120 ) = - //2, we obtain
The final wave functions should be normalized to complete the problem. Upon
normalization of Ib we obtain:
The two E functions from rps, 96, and cp7 can be obtained in yet another
way. First, one projects PE on 95 to obtain 29 5 - 96 - 97. Now, if one can
judiciously select for projection another function involving 96 and cp, (or sometimes
95 (6, and qp) that is orthogonal to the one just obtained, the desired function
will result. First, we try (6 and (p7 and realize that 96 -- (P will be orthogonal to
295 - (P6 - P7-
PE(96 - 7PO= 'PE96 - PE97 96 - 97
With these two examples, we see that the symmetry of the molecule determines
many aspects of our resulting wave functions. Next, we shall consider some
approximate molecular orbital calculations that complete the problem (but not
always necessarily correctly).
Molecular 3-6 HOCKEL PROCEDURE

Orbital The various ways to guesstimate values for the integrals presented in the
Calculations introduction correspond to various ways to do m.o. calculations. First, we shall
consider some very crude approximations, the so-called Huckel m.o. calculations,
which give a fair account of the properties of many hydrocarbons and which will
make the solution of the quantum mechanical problem in the introduction more
specific while the math remains simple. Generally,only the R orbitalsof hydrocarbon
systems are treated. The secular equations are written directly by generalizing
our earlier result for a two-atomic-orbital system as:
CI(HI I - ES1 1 ) + --- C,.(H1 . - ES1 .) = 0
C, 1(H.1 - ES. 1) + --- C(H,, - ESJ,) = 0
The following Hickel approximations are then introduced:

1. When two atoms i and j are not directly bonded, we take Hij = 0.
2. When all atoms are alike and if all bond distances are equal (e.g., the
carbon 2p, orbitals in benzene), all neighboring atom Hij values are taken as
equal and symbolized by #l (i.e., H 1 2 = H 2 3 = )-
3. If all atoms are alike, the Hi; integrals are symbolized by
a (jpft9i
1 dr = JP2 fi9 2 dT = 4L-
3-4 HOckel Procedure 69
4. Since we use normalized atomic orbitals in the basis set, S,, 1.
5. All Sjj's are set equal to zero.
The secular equations become
Cn(a - E) + C 12 # 12 ---- C 1j 1 , = 0
C.#,1 + - ---.-.- C,( - E) 0
The secular determinant now becomes
a -E #12 -.-.--.-.# 1n
=0 (3-22)
fln. ........ - E
where all but directly bonded fl's are zero. As a specific example, we shall consider
the it-system of the allyl radical in Fig. 3-5. (All sigma bonds are ignored.)
The 3 x 3 secular determinant can be directly written from equation (3-22) as:
C+
a-E # 0
#l -E # =0
0 #l X- E
Zeroes arise for the #313 matrix element, for the carbon atoms 1 and 3 are not
bonded. Dividing by # and letting (cx- E)/# = X, we obtain the determinant: FIGURE 3-5 The n-system of
the allyl radical.
X 1 0
1 X 1 =0
0 1 X
This determinant can be solved by expanding by the method of cofactors. The

procedure involves an expansion as shown for the general case:
a, b, c, b2 c2 a2 c a2 b2 =
a 2 b2 c2 =a1 -bb+ 0
aa3 2a
b3 C3
b3 C3 - a3 c3 a3 b3
The expansion can be done by row or column. In expanding by row, we take

each element separately and multiply it by the determinant that results when the
row and column containing the element are eliminated. Labeling the row i and
the column j, the sign is positive whenever i x j is odd and negative when i x j
is even.
For the allyl radical determinant, we expand to get:
X 1 1 1
X X
-1
0OX
-0
1
Solving these determinants gives X 3 - 2X = 0 with solutions X = 0, ± 2, i.e.,
a~ E
- = 0, 2, - 2.
These three roots produce the three energies:
Ei = a + F2#
E2 = 0 2
E, = oc- -,/2f
For the allyl radical, the secular equations are easily written by taking the
general set of secular equations and writing explicit ones for the three-atom allyl
7n system or by multiplying the column matrix {C 1 , C 2 , C3 } by the 3 x 3 secular
determinant:
C1X + C2 = 0 (3-23)
C1 + C2 X + C3 = 0 (3-24)
C2 + C3 X = 0 (3-25)
where X = (ca- E)/#.

For 01, the bonding molecular orbital, X = - 2, so substituting this into
(3-23) and (3-25) gives
- 2C 1 + C2 = 0 and C2 - 2C 3 = 0
Rearranging, we get:
C = 2 and C3 = C2
2
Using the normalization criterion:
C1 2 +C 2
2
+C 3 2 = 2 2 + C22 + (C2 2
1
2C 2 2 =1 and C 2 =-
Now
C2
C'=C3= C2
3-6 Huckel Procedure 71
and since
1
2
we obtain
1
C1 = C3e
2'
With the coefficients defined, we can write
11 1
1= ( I + 92 + (p3 (Bonding)
For 02
X =0 C 2 =0 and C1 = -C 3
1
C 12 + C 32 = 1 or C
\2
Accordingly, we can write
1 1
2 = 7 (P1 - 2 (p3 (Non-bonding)
Similarly
1 11
3 = 91 - 92 + (P (Antibonding)
The radical, cation, and anion of allyl are all described by adding the appropriate
number of electrons to these molecular orbitals.
Pictures of these orbitals can be drawn by realizing that a node exists every
time the wave function changes sign. In general, as the number of nodes in the
wave function increases, its energy increases. Furthermore, we see that all of the
symmetry information about the molecule is built into the molecular orbital
calculation. Using projection operators on the allyl radical, one can show that:
0 2 (A 2 ) = 2 ( -
P (p3)
1
0 1(B 1 ) = ((p1 + (p3) and 0 3 (B1 ) = (P2
72 (
The two B, m.o.'s have the same symmetry and can mix. Projection operators
cannot tell us the extent of mixing. Using a and b to describe the mixing, we can
write from symmetry the two wave functions:
ap 1 + b(p 2 + ap 3 and a'9 1 - b'P 2 + a'q 3
The Hfickel calculation gives the a's and b's, and also gives the same form for
the wave functions as the projection operator because all of the symmetry is
automatically included in the Hbckel Hamiltonian.
The symmetry considerations for the nitrite ion are the same as for allyl,
but the molecular orbital coefficients are different. Hfickel calculations have been
carried out on systems containing heteroatoms, X. These are treated by using
ax= ac + k/3. _
and
Pfc- = k' C
Tables of values to use for the quantities k and k' are listed in books on Huckel
m.o. calculations but will not be presented here because the general agreement
between fact and prediction deteriorates rapidly when Huckel calculations are
employed on molecules containing carbon bound to other kinds of atoms.
3-7 PROPERTIES DERIVED FROM WAVE FUNCTIONS
There are many quantities of chemical interest that can be calculated from the
wave functions. A few are listed and defined:
1. ELECTRON DENSITY AT AN ATOM r, q,

The electron density q, at an atom r in a molecule is given by:
q,= - njCr2 (3_26)
where nj is the number of electrons in the jth m.o. The summation is carried out
over all of the j occupied m.o.'s and CJr is the coefficient of the atomic orbital r
in the jth m.o.
Using the Hiickel wave functions for the filled molecular orbitals of the allyl
anion (two electrons in each 0, and 2 ):
1 1 1
1 1
2= 2 - 293
3-7 Properties Derived from Wave Functions 73
We sum (with n = 2) over the bonding and non-bonding m.o.'s and obtain for
the electron density on carbon 1:
q, = -2 - -2(=-
At carbon 2 we find:
q2 2( -)2(0) -1
The sum of all the q's equals the total number of electrons in the system.
2. FORMAL CHARGE, 6
The electron density at an atom in the molecule minus that on a free atom
is referred to as the formal charge on the atom in the molecule. Considering
carbon, we place one electron in each of the orbitals of the atom to form four
bonds. In Hickel theory, where we worry only about the 7r system, we consider
only the p. orbital on a carbon. Since it contains one electron, the charge on the
"free atom" is - 1. If the charge on this atom in a molecule (i.e., the electron
density at an atom r) is - 1.4, 6, the formal charge, is -0.4, that is, the difference
between the charge on this atom in the molecule and that on the neutral atom.
The sum of all the formal charges equals the total charge on the molecule.
For the allyl anion we obtain, using the charge density calculated above, a
charge of -0.5 at carbon 1 and a charge of zero at carbon 2.
3. BOND ORDER, PAB

The bond order between two adjacent atoms A and B is given by:
PAB fJCAjCBj
j (3-27)
i.e., just take the product of the coefficients of atoms A and B in each m.o. times
the number of electrons in the m.o. and sum over all of the m.o.'s.
Using this formula, the pi-bond order between carbon atoms I and 2 of the
allyl anion is given by
P12 = 2()(-) + 2( 0) 0.7
Huckel theory contains many crude approximations which can be justifiedm)

well for hydrocarbon molecules. With more polar molecules, they are very crude.
There have been many attempts to improve on these calculations. Extended
Hickel, CNDO, INDO, and others constitute various semi-empirical procedures
that have been used to approximate various integrals. In all of these procedures,
the overlap integral is evaluated and not set equal to zero. To go through the
various procedures in detail would take more time than we can allot to this topic
in this text. We will provide only a very rough overview of the calculations, and
will concern ourselves more with what can be done with the output.
3-8 EXTENDED HOCKEL PROCEDURE
In the extended HUckel method, the Hii's are approximated by valence state
ionization potentials. The Hij's, given by the Wolfsberg-Helmholz formula, are:
1
Hi = 1 kS..(H ; + Hjj) (3-28)
where k is a constant usually taken to be 1.8. All overlaps are explicitly evaluated.
Slater or Clementi atomic orbitals are generally used.
We shall briefly describe the criteria used to describe the electron density in
a molecule when S # 0 for i # j. With a LCAO wave function
0 1 = C1 <p
1 + C2 <P
2
we see that the total electron density, 0,2, is given by
f1
2
diT = C 1
2
2 dT + 2C 1 C 2 fP1c2 dT + C 2
2
f(P2 2dT
= C12 + 2C 1 C 2 S 12 + C 2 2
Using normalized a.o.'s (p12 = 1, we also see that jfp1<P2 dT is the overlap integral
S 12 . We shall now define some quantities that are important results of calculations
in which S # 0.
1. NET ATOMIC POPULATION

The quantity nfC 2 is referred to as the net atomic population (n, equals the
number of electrons in the m.o.). It is that part of the total electron density in
the m.o. that can be assigned to atom 1.
2. OVERLAP POPULATION
The quantity 2nC1 C 2 S 1 2 is referred to as the overlap population between

atomic orbitals 1 and 2 for one molecular orbital. In a polyatomic molecule, the
overlap population between atomic orbitals i and j is obtained by summing over
all occupied molecular orbitals k:
Zk 2nCCS,1 (3-29)
It is often convenient to have the atom-atom overlap population. This

corresponds to the total overlap of all the atomic orbitals on atom x with all
the atomic orbitals on atom y. Many m.o. programs present this as a reduced
3-8 Extended HUckel Procedure 75
overlap population matrix. This is
X
k(xy)
2nCjCjSij (3-30)
where k(xy) refers to the xy overlap population of every atomic orbital on x and
y summed over all filled molecular orbitals. When one compares bonds with
similar ionic character, the overlap population is related to the bond strength.
3. GROSS ATOMIC POPULATION

This is a procedure for assigning all the electron density in a molecule to
each of the individual atoms. Half of the bonding electron density (e.g., half of
2nC 1C2 Su2 ) is given to each of the atoms involved in bonding.
Gross atomic population = X (nC7 + nCjCjSjj) (3-31)

all
m.o. s
4. FORMAL CHARGE
The formal charge on an atom is equal to the core charge (nucleus plus all
non-valence electrons) minus the gross atomic population.
To give you some appreciation for the kind of information available from
a semi-empirical molecular orbital calculation, the output from an extended
Hilckel m.o. calculation is presented in Table 3-1. The matrix summarizes the
wave function; e.g., from m.o. 12, we see that
2 = -0.40 C(2)2p, - 0.56 C(1)2p - 0.40 C(3)2p,
(Numbers in parentheses refer to the numbering in the figure of allyl on the

second page of computer output.) This is the lowest-energy filled pi-orbital. The
odd electron in m.o. 9 is delocalized over both of the terminal carbons, and these
are the only two orbitals contributing to this molecular orbital.
One important concept to be gained from this calculation is the extent of
delocalization of the bonding in the sigma system, as can be seen by looking at
a sigma m.o. For example, none of the filled m.o.'s even vaguely resemble a
localized carbon-hydrogen sigma bond. All of the a m.o.'s are very extensively
delocalized. The nr-type atomic orbitals are orthogonal to the a m.o.'s, and they
contain only contributions from the carbon p, orbitals. Net charges, orbital
occupations, and an atom-by-atom overlap population are evaluated in this
program from the wave functions calculated.
Extended HUckel calculations give reasonable wave functions for systems
in which the formal charges on the atoms of the molecule are small, + 0.5. As
an example of the utility of such calculations, the results from a study of several
metallocenes(2 ) and bis-arene complexes(3 ) will be discussed. In all systems studied,
the correct ground state results; i.e., the unpaired electrons end up being placed
in molecular orbitals that are essentially the correct d orbitals. Certain ground
state spectroscopic properties (nmr and epr) are also predicted. This agreement
with experiment suggests that the wave functions are reasonably good. They are
TABLE 3-1 Wave Function from Extended HOckel Calculation of the Allyl
Radical
(The vertical numbers correspond to A.O.'s and the horizontal ones to M.O.'s)
A.O.'s M.O.'s
1 2 3 4 5 6 7 8, 9,
1 -1.80 -0.00 0.24 -0.00 0.27 -0.00 0.25 -0.00 0.00
2 0.00 0.00 0.00 -0.00 0.00 -0.00 0.00 -0.97 -0.00

3 -0.51 -0.00 -0.35 0.00 -1.31 0.00 -0.24 0.00 -0.00
4 0.00 -1.28 -0.00 -0.68 -0.00 -0.29 -0.00 -0.00 -0.00

5 0.76 1.12 0.87 -0.38 -0.01 0.22 0.01 -0.00 0.00
6 -0.00 -0.00 -0.00 0.00 0.00 -0.00 0.00 0.65 -0.74

7 -0.43 -0.53 0.31 -0.58 0.30 0.62 -0.65 0.00 -0.00
8 -0.68 -0.13 0.58 -0.89 -0.53 -0.54 0.15 -0.00 0.00

9 0.76 -1.12 0.87 0.33 -0.01 -0.22 0.01 -0.00 -0.00
10 -0.00 -0.00 -0.00 0.00 -0.00 0.00 -0.00 0.65 0.74

11 -0.43 0.53 0.31 0.58 0.30 -0.62 -0.65 -0.00 -0.00
12 0.68 -0.13 -0.58 -0.89 0.53 -0.54 -0.15 0.00 0.00

13 0.28 0.00 -0.30 0.00 -1.17 0.00 -0.61 -0.00 0.00
14 -0.07 -0.48 -0.61 0.49 0.46 0.68 -0.59 0.00 0.00

15 0.15 -0.09 -0.75 0.82 -0.03 -0.55 0.61 -0.00 -0.00
16 0.15 0.09 -0.75 -0.82 -0.03 0.55 0.161 0.00 0.00

17 -0.07 0.48 -0.61 -0.49 0.46 -0.68 -0.59 0.00 -0.00
10 11 12, 13 14 15 16 17
1I -0.00 0.01 -0.00 0.00 -0.09 0.33 -0.00 0.48
2 -0.00 -0.00 -0.56 0.00 -0.00 -0.00 -0.00 0.00

Ci 3 -0.00 0.33 -0.00 -0.00 -0.36 -0.17 -0.00 0.01
4 -0.49 -0.00 -0.00 -0.30 0.00 -0.00 0.16 -0.00

5 -0.14 0.05 0.00 -0.03 -0.03 -0.20 0.44 0.36
6 -0.00 0.00 -0.40 -0.00 0.00 -0.00 0.00 0.00

C 7 -0.12 -0.35 -0.00 0.37 -0.12 -0.12 0.01 -0.00
8 0.41 0.00 -0.00 0.07 0.29 -0.14 0.02 -0.01

9 0.14 0.05 -0.00 0.03 -0.03 -0.20 -0.44 0.36
10 0.00 0.00 -0.40 0.00 -0.00 -0.00 0.00 0.00

C 11 0.12 -0.35 -0.00 -0.37 -0.12 -0.12 -0.01 -0.00
12 0.41 -0.00 -0.00 0.07 -0.29 0.14 0.02 0.01

3-8 Extended Huckel Procedure 77
TABLE 3-1 (continued)
13 0.00 -0.40 -0.00 0.00 0.27 0.32 0.00 0.08
14 0.37 0.26 0.00 -0.15 0.27 -0.15 0.17 0.04

H 115 -0.07 -0.33 0.00 0.36 0.02 -0.23 0.19 0.05
16 0.07 -0.33 0.00 -0.36 0.02 -0.23 -0.19 0.05

17 -0.37 0.26 -0.00 0.15 0.27 -0.15 -0.17 0.04
M.O.'s 10 to 17 have 2e-
9 has le-
lto 8 empty
S
Atomic Orbitals 1-4, 5-8, and 9-12 are carbon orbitals in the order 2s, 2pz, 2 p, 2p, for each atom.
A.O.'s 13 to 17 are hydrogen Is. Gross Atomic Population
Atomic
M.O. Energy (eV) Occ. Orbital Occupations
1 53.22 0
2 36.55 0
3 28.31 0
4 22.97 0
5 18.76 0 5 1.17
6 7.30 0 6 1.01
7 5.19 0 7 0.96
8 -4.63 0 8 0.89
9 -10.16 1 9 1.17
10 -12.67 2 10 1.01
11 -13.83 2 11 0.96
12 -14.03 2 12 0.89
13 -15.01 2 13 0.98
14 -15.96 2
15 -19.40 2
16 -22.83 2
17 -27.54 2
Coordinate
System
Atom
V Number Net Charges
H(5)
C H 1 0.013
(6) 2 -0.020
(2)
3 -0.030
H---C )-----------X
4 0.015
(4)
5 0.000
C H 6 0.016
/(3) (7)
H 7 0.017
8 0.000
(8)
(Continued)
Reduced Overlap Population Matrix Atom by Atom
1 2 3 4 5 6 7 8
1 5.236 1.411 1.141 0.836 -0.117 -0.075 -0.075 -0.117
2 1.141 5.604 -0.216 -0.097 0.821 0.812 -0.013 0.009

3 1.141 -0.216 5.604 -0.097 0.009 -0.013 0.812 0.821
4 0.836 -0.097 -0.097 1.360 -0.020 0.005 0.005 -0.020

5 -0.117 0.821 0.009 -0.020 1.402 -0.094 0.000 -0.000
6 -0.075 0.812 -0.013 0.005 -0.094 1.334 -0.000 0.000

7 -0.075 -0.013 0.812 0.005 0.000 -0.000 1.334 -0.094
8 -0.117 0.009 0.821 -0.020 -0.000 0.000 -0.094 1.402
then analyzed in detail to provide information about the bonding interactions

responsible for bonding the hydrocarbon ring to the metal. An interesting result
involves the extensive mixing of the ring sigma orbitals with the metal orbitals
in these systems.
3-9 SCF-INDO (INTERMEDIATE NEGLECT OF

DIFFERENTIAL OVERLAP)
Before discussing the INDO calculation, it is appropriate to discuss in more

detail some of the approximations that have been employed without mention in
the methods discussed for many-electron systems. The wave function will be
considered first. The ground state of our molecule is described by a wave function
T, which, when substituted into equation (3-4), would give us the energy, if the
equation could be rigorously solved. In the LCAO approach, we are attempting
to represent T by a combination of one-electron functions (i.e., molecular orbitals),
each of which depends only on the coordinates of one electron. For an n-electron
system, the ground state wave function is given by simply taking the product of
the occupied one-electron orbitals, i.e.,
P(1, 2 ... n) = t 1(1)0 2 (2) ... - (n) (3-32)
The subscript on 0 labels the orbital used and the number in parentheses describes
the electron. Physically, this amounts to making the untenable assumption that
electron 1 is independent of the n - 1 other electrons in the system. We shall
have to correct or pay the consequences for this later.
If the ft's were written as the sum of one-electron operators, the Schrodinger
equation could be solved by a simple separation of variables. This would amount
to ignoring completely the existence of the other electrons; e.g., the e 2 /rij electron
repulsion terms would not be included. What one does is to use one-electron
operators, and correct for the potential experienced at an electron p from the
field of the atomic nuclei and the average field of the other n - 1 electrons. We
label this field as V,,, and it can be approximated as will be discussed shortly.
Essentially, then, our Hamiltonian is R(,), given by
R(P) = Y F(,) = 2m () 2 + ,P) (3-33)
When R(,, is used in a Schr6dinger-type equation, we have:
R,,,(1, 2 - n)P(1, 2 - n) = & (1, 2 -..)
Note that our operator R(,) can be separated (separation of variables) into a sum
of one-electron operators F(,), which give rise to a set of equations of the form
FM =-() Eoifr(1) (3-34)
The Orfunctions are the one-electron molecular orbitals obtained by the variation
procedure. These are the wave functions we have been talking about when we
do Hickel or extended Huckel calculations. In these cases, all of the electron-
electron repulsions are very indirectly lumpect into the parameters employed, the
charge correction and the Wolfsberg-Helmholtz type of approximation. The
Hamiltonian is not explicitly defined.
In self-consistent field (SCF) methods, the Hamiltonian is explicitly written
and used. The VP term of equation (3-33) is an e2 r operator. In an INDO
calculation, the formalism is basically SCF (vide infra), but many of the resulting
integrals are ignored. In SCF calculations, the Hamiltonian for an n-electron
problem is written as:
e2
= f i+ I e (3-35)
i=1 i<j rij
Here, rij is the interelectronic distance and the i < j index ensures that electron-
electron interactions are counted only once. Ai, the operator for electron i in the
field of the nuclei, consists of a kinetic energy term and an electron-nuclear
attraction term:
h2 , e2
Hi= V2 - e(3-36)
2m k rik
where Vi2 is the familiar kinetic energy operator for electron i, rik is the
electron-nucleus distance, and Zk is the charge on the kth nucleus. The energy
is found by a variational procedure, and successive sets of trial wave functions
are used until this energy is minimized. These calculations are made more difficult
because the operators depend on the wave functions. That is, in order to evaluate
e2/rij we must know the electron distribution, so we must know the wave function
before we calculate the wave function. A starting wave function is guessed, the
calculations are carried out, and a new wave function is obtained. The new wave
function is then used to repeat the calculation. The procedure is iterated until
there are no further changes in the wave functions, i.e., we have a self-consistent
field. This still is not a rigorous procedure because we are treating the electrons
in a molecule as if they experience a charge cloud from all the other electrons.
In actual fact, the electron motion is correlated; i.e., the motion of one electron
is governed by the instantaneous positions of the other electrons and not their
average. This effect, called electron correlation, is often ignored, leading to
erroneous wave functions.
In any molecule of more than four or five atoms, even a limited basis set
(i.e., one in which only the valence atomic orbitals on each atom are employed)
gives rise to a huge number of integrals when all interactions (kinetic energy,
nuclear-electron, nuclear-nuclear, and electron-electron) are taken into account.
The various calculation procedures CNDO (complete neglect of differential
overlap) CNDO-2, INDO (intermediate neglect of differential overlap), and so
forth, differ in the types of integrals that are ignored. At present, INDO is the
most popular of these approximate methods.
We can illustrate the types of integrals encountered by considering the
following four orbitals taken from a large molecule:
i
4- k
Electron repulsion integrals of the form
ftr,
f i*(1)cp *( 1 pk(2 )pi( 2 ) dT, dr2
are the most numerous and require the most computer time to integrate
numerically. This integral represents the 1/r repulsion between the electron density
in the region where (pi and cp have differential overlap and the electron density
in the region where 9 , and (p, have differential overlap. The differential overlap
for electron 1, for example, is simply defined as
j(1)o (1) drl
It is important to emphasize that although an overlap integral may be zero,
Ttk(01)(1) dz, = 0
an integral such as
f6 p dr,
,k(l)q1(1)
TABLE 3-2. INDO Calculation for Allyl

(One atomic unit of energy is defined as twice the energy of the first Bohr orbit in the hydrogen
atom; i.e., twice the ionization potential of a hydrogen atom.)
Electronic Energy 60.0089
Energy Satisfied
Eigenvalues and Eigenvectors
eV - -52.19 -42.62 -36.32 -33.74 -30.21 -26.56 -23.36 -23.17

Eigenvalues A.U. - -1.915 -1.566 -1.335 -1.240 -1.110 -0.976 -0.859 -0.851
(orbital energies)
1 2 3 4 5 6 7 8
1 1 C S -0.62 -0.00 -0.34 0.02 0.00 0.00 0.00 -0.10

2 1 C Px -0.13 -0.00 0.29 0.48 -0.00 -0.00 -0.00 0.40
3 1 C Py -0.00 0.42 -0.00 0.00 -0.45 -0.00 0.33 0.00

4 1 C Pz -0.00 0.00 0.00 -0.00 0.00 -0.72 0.00 -0.00
5 2 C S -0.45 0.51 0.27 -0.06 0.03 -0.00 0.16 0.06

6 2 C Px 0.08 -0.08 0.30 0.20 0.44 0.00 0.24 -0.36
7 2 C Py 0.19 0.11 0.21 -0.43 0.25 0.00 -0.42 0.02

8 2 C Pz 0.00 0.00 -0.00 -0.00 -0.00 -0.49 -0.00 -0.00
9 3 C S -0.45 -0.51 0.27 -0.06 -0.03 -0.00 -0.16 0.06

10 3 C Px 0.08 0.08 0.30 0.20 -0.44 0.00 -0.24 -0.36
11 3 C Py -0.19 0.11 -0.22 0.43 0.25 -0.00 -0.42 -0.02

12 3 C Pz 0.00 -0.00 -0.00 -0.00 0.00 -0.49 0.00 -0.00
13 4 H S -0.20 -0.00 -0.33 -0.28 0.00 0.00 0.00 -0.47

14 5 H S -0.13 0.30 0.15 0.33 0.00 -0.00 -0.40 0.27
15 6 H S -0.13 0.23 0.34 -0.05 0.38 0.00 0.17 -0.31

16 7 H S -0.13 0.23 0.34 -0.05 -0.38 0.00 -0.17 -0.31
17 8 H S -0.13 0.30 0.15 -0.33 -0.00 0.00 0.40 0.27

-8.53 -7.05 -7.07 -0.08 -0.20 0.81 1.79 4.48 5.76

Elgenvalues - -0.314 -0.068 -0.051 -0.033 -0.029 0.030 0.066 0.165 0.212
9 10 11 12 13 14 15 16 17
1 1 C S 0.00 0.43 0.00 -0.19 0.00 0.00 0.43 0.30 -0.00

2 1 C Px 0.00 -0.09 -0.00 -0.20 0.00 0.00 -0.31 0.60 -0.00
3 1 C Py -0.00 0.00 -0.29 -0.00 0.00 -0.12 0.00 0.00 0.65

4 1 C Pz 0.00 -0.00 0.00 0.00 0.69 0.00 -0.00 0.00 0.00
5 2 C S 0.00 -0.26 0.35 0.37 -0.00 0.25 0.07 -0.10 -0.18

6 2 C Px 0.00 -0.22 0.29 -0.24 0.00 -0.23 0.37 0.02 0.32
7 2 C Py -0.00 0.08 -0.08 0.14 -0.00 0.37 0.16 0.42 0.34

8 2 C Pz 0.71 0.00 -0.00 -0.00 -0.51 0.00 0.00 -0.00 0.00
9 3 C S -0.00 -0.26 -0.35 0.37 -0.00 -0.25 0.07 -0.01 0.18

10 3 C Px -0.00 -0.22 -0.29 -0.24 0.00 0.23 0.36 0.02 -0.32
11 3 C Py -0.00 -0.08 -0.08 -0.14 0.00 0.37 -0.16 -0.42 0.34

12 3 C Pz -0.71 0.00 0.00 -0.00 -0.51 -0.00 0.00 -0.00 -0.00
13 4 H S -0.00 -0.50 -0.00 -0.16 0.00 0.00 -0.43 0.33 -0.00

14 5 H S 0.00 -0.08 0.08 -0.48 0.00 -0.49 0.01 -0.16 -0.04
15 6 H S -0.00 0.39 -0.49 -0.04 0.00 -0.06 -0.33 -0.08 -0.19

16 7 H S 0.00 0.39 0.49 -0.04 -0.00 0.06 -0.33 -0.08 0.19
17 8 H S -0.00 -0.08 -0.08 -0.48 0.00 0.49 0.01 -0.19 0.04
Binding Energy = - 2.6884 A.U. = 73.15 eV Dipole Moment = 3.09 Debyes
is not necessarily zero. For example, when the operator 6, is 1/r1N, we have the
electron-nuclear attraction integral, which cannot be assumed to be zero.
Note that p, and p, overlap very little, so the differential overlap
j 2
,P( )(p,( 2 )dT 2 is almost zero everywhere. The value of the integral is thus very
small. If a systematic method were developed to set certain of these differential
overlaps to zero, the number of such three- and four-center integrals would be
drastically reduced. This is just what the various NDO schemes do.
For purposes of comparison with the results from the Hckel calculation
and the extended Huckel output, the computer output from the INDO calculation
on allyl is presented in Table 3-2. The eigenvalues of the energy are given above
each molecular orbital in units of electron volts and atomic units.
It is of interest to examine 0, in the INDO output:
fL9 = 0.71((p - (12)
It is normalized, since
f f 92 dT = (0.71)2 If ps dz - 2 f8912 dr + j' 12P

12dT]
3-10 Some Predictions from M.O. Theory 83
=(0.7 1)2(1 - 0 + 1)
=-1
Now look at 0, in the extended Hickel calculation:
0 9 = 0.74(q 10 - (6)
It is the same orbital with the same symmetry, but it has a different coefficient.
Is it still normalized? Yes, because extended Hiickel does not neglect differential
overlaps, so
T910 (P6 dc = 0
and the normalization procedure as used above will yield a different coefficient.
Of the various INDO programs available, the one referred(') to as INDO/1
or ZINDO is particularly good for reproducing geometries of inorganic sys-
tems.(') The method has been used(6) to calculate points on the potential energy
surface of alkane hydroxylations and alkene epoxidations by metal oxo species.
The results lead to reaction mechanisms consistent with those proposed from
experimental observations.
3-10 SOME PREDICTIONS FROM M.O. THEORY ON

ALTERNATELY DOUBLE BONDED HYDROCARBONS
You probably have seen examples of the successes of m.o. theory for homonuclear
diatomics, e.g., the prediction of paramagnetic 02. One further success involves
the prediction of the structures of even numbered, alternately double bonded
hydrocarbons. The systems C4H 4 , CH, and C16111 are more stable as a set
of alternating double bonds than they are as a delocalized pi-system. On the
other hand, C6 H, C,,H,,, C 14 1 4 , and Cis1Hi1 (i.e.,, 4n + 2, when n = 1, 2, 3,
... ) are aromatic.
The m.o. results for these even numbered, alternately double bonded
hydrocarbons show a singly degenerate low energy orbital and a singly degenerate
high energy level with doubly degenerate levels in between:
C4 C, C8 C1 0 C12 C1 4 etc.
E
* 4F*
+-+
**i
*~1 +1
Note that the molecules with two unpaired electrons are stabilized much less
than the other molecules because the highest energy electrons are in non-bonding
m.o.'s. These molecules will in fact be more stable if they undergo a distortion
in which the delocalized molecular orbital is converted into two localized double
bonds, as shown below:
++ - ------
Delocalized Two
4-membered ring double bonds
3-11 MORE ON PRODUCT GROUND STATE WAVE

FUNCTIONS
The product wave function written in equation (3-32) is not complete. If we write
a wave function, interchange of the labels on the electrons 1, 2, ... n cannot
change any physical property associated with the electron density, V2, of the
system. In order for T 2 to be unaffected, the effect of the interchange on T must
produce + so that T 2 is unchanged. To demonstrate the problem, we shall
define a permutation operator which switches the electron labels. For a
two-electron system 2 and # in orbital 0 1, we can write
T = j1(1)i(1)0 1 (2)#l(2) (3-37)
The permutation operator switches the electron labels in parentheses to
P 12 1(1, 2) = 0 1(2)2x(2) 1 (1)#l(1) (3-38)
The result is not related to our initial wave function by ± 1.

However, consider the wave function
T = 01(1)i(1)01(2)#l(2) - 0j(2)2c(2)0,(1)#(1) (3-39)
Now
P12T = 0 1(2)2(2)0 1(1)#(1) - 1 (1)2(1)@ 1(2)#(2) (3-40)
This result (equation (3-40)) is related to our starting wave function by -1. Since
we have the result P'T = -1T, we have an antisymmetrized wave function. Had
we chosen
T = 0 1 (1)o(1) 1 (2)f#(2) + 1 (2)a(2) 1(1)fl(1)
the result would have been + 1. This is a symmetrized wave function. Later we
shall show that the antisymmetrized function must be chosen in order that the
Pauli exclusion principle apply.
References 85
One can write the correct antisymmetrized wave function for a 2n-electron
system by constructing a Slater determinant and expanding it. The determinant
is constructed as follows:
1(1)fl(l)f2(1)2(1) . -f.(1)#(1)
T = N 0 1(2),(2) 0 1(2)P(2) -........
0 1(2n)c(2n) - - -- l,(2n)f#(2n)
Expanding this determinant by the method of cofactors produces the antisym-
metrized wave function. Try it for the two-electron system just discussed.
The properties of determinants are consistent with several properties of
electrons in molecules. For example, if rows of a determinant are changed, so is
the sign. Exchanging rows corresponds to the use of the permutation operator,
e.g., equation (3-40). Thus, the antisymmetry property is built into the determi-
nant. Further, there is a theorem stating that if two columns of a determinant
are identical, the determinant vanishes. Thus, a non-zero wave function cannot
be constructed if two electrons with the same spin are placed in the same orbital.
1. M. J. S. Dewar, "The Molecular Orbital Theory of Organic Chemistry," McGraw- REFERENCES

Hill, New York, Chap. 5, 1969. CITED
2. M. F. Rettig and R. S. Drago, J. Amer. Chem. Soc., 91, 3432 (1969).
3. S. E. Anderson, Jr., and R. S. Drago, Inorg. Chem., 11, 1564 (1972).
4. W. Anderson, W. D. Edwards, and M. C. Zerner, Inorg. Chem., 25, 2728 (1986) and
references therein. The program is available by contacting M. C. Zerner.
5. a. T. R. Cundari, M. C. Zerner, and R. S. Drago, Inorg. Chem., 27, 4239 (1988).
b. W. P. Anderson, T. R. Cundari, R. S. Drago, and M. C. Zerner, Inorg. Chem., 29, 1
(1990).
6. a. T. R. Cundari and R. S. Drago, Intern. J. Quantum Chem., 24, 665 (1990).
b. T. R. Cundari and R. S. Drago, Inorg. Chem., 29, 487 (1990).
M. J. S. Dewar, "The Molecular Orbital Theory of Organic Chemistry," McGraw-Hill, ADDITIONAL

New York, 1969. READING
J. M. Anderson, "Introduction to Quantum Chemistry," W. A. Benjamin, New York, 1969.
A. Streitwieser, "Molecular Orbital Theory for Organic Chemists," Wiley, New York, 1961.
A. Liberles, "Introduction to Molecular Orbital Theory," Holt, Rinehart and Winston,
New York, 1966. (A simple approach to Hiickel theory with many worked examples.)
W. Kauzman, "Quantum Chemistry," Academic Press, New York, 1957.
J. A. Pople and D. L. Beveridge, "Approximate Molecular Orbital Theory," McGraw-Hill,
New York, 1970.
W. G. Richards and J. A. Horsley, "Ab Initio Molecular Orbital Calculations for Chemists,"
Oxford University Press, New York, 1970.
COMPILATIONS W. G. Richards, T. E. H. Walker, and R. K. Hinkley, "A Bibliography of Ab Initio

Molecular Wave Functions," Oxford University Press, New York, 1971.
W. G. Richards et al., "Bibliography of Ab Initio Molecular Wave Functions-Supplement
for 1970-73," Oxford University Press, New York, 1974.
A. Hammett, Specialist Periodical Reports, The Chemical Society, London, England.
H. Hulbronner and C. Straub, "Hickel Molecular Orbitals, Springer-Verlag, New York,
1966.
A. Streitwieser and J. Brauman, "Supplemental Tables of Molecular Orbital Calculations,"
Volumes I and II, Pergamon Press, New York, 1965.
EXERCISES 1. Using the Huckel wave functions reported in this chapter, calculate the formal charges
on the carbons and the C-C bond orders for
a. the allyl cation
b. the allyl radical
2. a. Using projection operators, work out the symmetry-adapted linear combinations

of pi orbitals for the allyl radical.
b. Use these linear combinations to obtain a set of orthonormal molecular orbitals.
c. Sketch the shapes of the 7r and 7r* molecular orbitals.
d. Compare your results with those obtained from the Huckel calculation described
in this chapter.
3. a. Using projection operators, work out the symmetry-adapted linear combinations

of pi orbitals for cyclobutadiene.
b. Use these linear combinations to obtain a set of orthonormal molecular orbitals.
c. Sketch the shapes of the 7n and 7r* molecular orbitals.
d. Would any more information be obtained from a Huckel calculation? If so, what?
4. For the 7n system of borazole, B3N 3 H 6 , obtain:
(1)
B
(6)N N (2)
(5) B N B (3)
(4)
a. the total representation

Exercises 87
b. all of the wave functions
c. the normalization factor for each wave function; and indicate which wave functions
can mix.
5. Propargylene is a linear diradical with the following valence bond structure:
H-C=C-C--H -+ Z-axis
2
Using the three 2p, and three p, orbitals of the carbons as a basis set,
a. How many m.o.'s will be produced?
b. Do a simple Huckel calculation to determine the energies of the m.o.'s from part a.
c. What are the a.o. coefficients of the two largest energy m.o.'s?
d. Calculate the bond order between two adjacent carbons.
6. The ion Re 2 C18 2- is made up of two square planar ReCl4 units joined by a Re-Re
bond. The two planes are parallel and the chlorines are eclipsed (as viewed down the
Re-Re bond, which we define as the z-axis). The symmetry is therefore D4 .
a. Determine the symmetries of the m.o.'s obtained using the two Re d,2 orbitals as
the basis set. Use projection operators to determine these m.o.'s and sketch them.
b. Repeat part a using the two dx,2 -2 orbitals as the basis set. These orbitals point
directly at the chlorines.
c. Applying this same procedure to the d, orbitals, one obtains

x
p(b2 g) - (d(')+ dm
b) - 2 (d -d + +
Re Re
Fo r d , an d d3, : +d + +
01(e.)=1)
02(e.) = -(dj') +
+ d, 22)
) (
/2
2
0 2(e) =- (d() - d, ))
Sketch a plausible energy level diagram. Fill in the electrons and determine the
Re-Re bond order. (Hint: the bonding orbital obtained in part b participates
in dsp 2 hybrids that are filled by chlorine electron pairs. The other orbitals must be
filled by Re electrons.)
d. Ignore the chlorines and treat the ion as if it belonged to Do. What are the
symmetries of the 10 m.o.'s obtained above? (Note: this illustrates the origin of the
terms --bond, it-bond, and so forth.)
H
7. Consider the H3 + molecule in both the linear (H-H-H)* and triangular ( /\ )*
states. H H
a. Utilizing simple Huckel m.o. theory with the hydrogen Is orbitals as a basis set, set
up the secular determinant and compute the energies of the m.o.'s for each geometry.
b. From your calculation, predict which geometry would be most stable for H3 * and
for H 3 -
c. Consider only the linear geometry again, using the hydrogen Is orbitals as a basis
set. How many molecular orbitals will be formed? Substitute the energies found for
the m.o.'s in part a into the secular determinant and find the orthonormal set of
LCAO m.o.'s.
8. a. Two possible structures for the ion 13 are
[I-I-iI
Using a basis set of p. and p, orbitals, simple Huckel molecular orbitals can be
calculated for the two structures. Ignoring any it bonding, calculate the energies of
the three sigma m.o.'s for the linear case, using the three atomic p2 orbitals shown
above. (Do not calculate coefficients.)
b. Making the assumptions that # = j(P;Sbeeteg = j(Piiineaj,9 and a f(Pjfeent'Pj

JPjHineapj, which structure is more stable if the total energy of the
p-orbital lone pairs and the m.o.'s in the bent configuration is 16C + 2#3? (Hint:
There will be six non-bonding p-orbital lone pairs, which will contribute 120t to the
total energy of the linear structure.)
Exercises 89
9. Consider the trigonal bipyramidal CuCl5 3- anion:
Cl
Cl Cl
SCl
Choose axes at the Cl- ligands such that z points upward, parallel to C3. For the
in-plane Cl- nuclei, let x point toward the Cu2 . Thus, from the bottom the xy-system
for the ligands looks like the following:
y
x
yy
Choose the x and y axes as you wish for the axial Cl ligands.
a. Find the total representation of the p orbitals of the five chlorines and decompose it
into irreducible representations.
b. Use a priori reasoning to determine the effect on an equatorial p, or p, orbital of the

projection operator Pg , where x" denotes any double primed irreducible representa-
tion. (Hint: What do these irreducible representations have in common?)
c. Operate on an equatorial p, with PE' to find a symmetry-adapted linear COmbination

with E' symmetry.
General Introduction
to Spectroscopy
4-1 NATURE OF RADIATION
There are many apparently different forms of radiation, e.g., visible light, radio
waves, infrared, x-rays, and gamma rays. According to the wave model, all of
these kinds of radiation may be described as oscillating electric and magnetic
fields. Radiation, traveling in the z-direction for example, consists of electric and
magnetic fields perpendicular to each other and to the z-direction. These fields
are represented in Fig. 4-1 for plane-polarized radiation. Polarized radiation was
selected for simplicity of representation, since all other components of the electric
field except those in the x-z-plane have been filtered out. The wave travels in the
z-direction with the velocity of light, c(3 x 1010 cm sec-'). The intensity of the
radiation is proportional to the ampltiude of the wave given by the projection
on the x- and y-axes. At any given time, the wave has different electric and
magnetic field strengths at different points along the z-axis. The wavelength, 2,
of the radiation is indicated in Fig. 4-1, and the variation in the magnitude of
this quantity accounts for the apparently different forms of radiation listed above.
If the radiation consists of only one wavelength, it is said to be monochromatic.
Polychromatic radiation can be separated into essentially monochromatic beams.
For visible, u.v., or IR radiation, prisms or gratings are employed for this purpose.
Radiation consists of energy packets called photons, which travel with the
velocity of light. The different forms of radiation have different energies.
x
V
FIGURE 4-1 Electric and

magnetic field components of
plane-polarized
electromagnetic radiation.
4-2 Energies Corresponding to Various Kinds of Radiation 91
In our discussion of rotational, vibrational, and electronic spectroscopy, our

concern will be with the interaction of the electric field component of radiation
with the molecular system. This interaction results in the absorption of radiation
by the molecule. In epr and nmr the concern is with the interaction with the
magnetic component of radiation.
In order for absorption to occur, the energy of the radiation must match
the energy difference between the quantized energy levels that correspond to
different states of the molecule. If the energy difference between two of these
states is represented by AE, the wavelength of the radiation, 2, necessary for
matching is given by the equation:
AE = hc/2 or A=hc/AE (4-1)
where h is Planck's constant, 6.623 x 10-27 erg sec molecule ', and c is the
speed of light in cm sec', giving AE in units of erg molecule-'. Equation (4-1)
relates the wave and corpuscular models for radiation. Absorption of one quantum
of energy, hc/2, will raise one molecule to the higher energy state.
As indicated by equation (4-1), the different forms of electromagnetic
radiation (i.e., different 2) differ in energy. By considering the energies corre-
sponding to various kinds of radiation and comparing these with the energies
corresponding to the different changes in the state that a molecule can undergo,
an appreciation can be obtained for the different kinds of spectroscopic methods.
4-2 ENERGIES CORRESPONDING TO VARIOUS KINDS

OF RADIATION
Radiation can be characterized by its wavelength, 2, its wave number, v, or its

frequency, v. The relationship between these quantities is given by equations
(4-2a) and (4-2b):
v(sec -) c(cm sec') (4-2a)

2(cm)
1
v(cm -') = 1 (4-2b)
(cm)
The quantity v has units of reciprocal centimeters, for which the official IUPAC
nomenclature is a kaiser; 1000 cm- are equal to a kilokaiser (kK). From
equations (4-1) and (4-2), the relationship of energy to frequency, wave number,
and wavelength is:
AE(ergs molecule -') = hv = hc/2 = hc (4-3)
In describing an absorption band, one commonly finds several different units

being employed by different authors. Wave numbers, i, which are most commonly
employed, have units of cm - and are defined by equation (4-2b). Various units
are employed for 2. These are related as follows: 1 cm = 108 A (angstroms) = 107
nanometers = 10' p (microns) = 107 my (millimicrons). The relationship to vari-
ous common energy units is given by: 1 cm-' = 2.858 cal/mole of par-
92 Chapter 4 General Introduction to Spectroscopy
ticles = 1.986 x 10 - erg/molecule = 1.24 x 10 - eV/mole. These conversion

- (cm-1) units can be employed to relate energy and wavelength; or the equation
Radio
frequency - 10 2
AE(kcal mole-') x 1(A) = 2.858 x 105 (4-4)
- 10-1
Microwave
can be derived to simplify the calculation of energy from wavelength.
Wave numbers corresponding to various types of radiation are indicated in
Fig. 4-2. The small region of the total spectrum occupied by the visible portion
-- 10
Far
is demonstrated by this figure. The higher energy radiation has the smaller
IR - 102 wavelength and the larger frequency and wave number (equation (4-3)). The
following sequence represents decreasing energy:
Near - 103
IR
- 104 ultraviolet > visible > infrared > microwave > radio-frequency
vis
NearUV
Far
uv 106 4-3 ATOMIC AND MOLECULAR TRANSITIONS
Xray - 107
In an atom, the change in state induced by the quantized absorption of radiation
--- 108 can be regarded as the excitation of an electron from one energy state to another.
y ray The change in state is from the ground state to an excited state. In most cases,
the energy required for such excitation is in the range from 60 to 150 kcal mole - .
Calculation employing equation (4-3) readily shows that radiation in the
FIGURE 4-2 Wave numbers
ultraviolet and visible regions will be involved. Atomic spectra are often examined
of various types of radiation.
as emission spectra. Electrons are excited to higher states by thermal or electrical
energy, and the energy emitted as the atoms return to the ground state is measured.
In molecular spectroscopy, absorption of energy is usually measured. Our
concern will be with three types of molecular transitions induced by electromag-
netic radiation: electronic, vibrational, and rotational. A change in electronic
state of a molecule occurs when a bonding or nonbonding electron of the molecule
in the ground state is excited into a higher-energy empty molecular orbital. For
example, an electron in a it-bonding orbital or a carbonyl group can be excited
into a nt* orbital, producing an excited state with configuration oJlt'7*l. The
electron distributions in the two states (ground and excited) involved in an
electronic transition are different.
The vibrationalenergy states are characterized by the directions, frequencies,
and amplitudes of the motions that the atoms in a molecule undergo. As an
example, two different kinds of vibrations for the SO2 molecule are illustrated
in Fig. 4-3. The atoms in the molecule vibrate (relative to their center of mass)
in the directions indicated by the arrows, and the two extremes in each vibrational
mode are indicated. In the vibration indicated in (A), the sulfur-oxygen bond
length is varying, and this is referred to as the stretching vibration. In (B), the
motion is perpendicular to the bond axis and the bond length is essentially
constant. This is referred to as a bending vibration. In these vibrations, the net
effect of all atomic motion is to preserve the center of mass of the molecule so
that there will be no net translational motion. The vibrations indicated in Fig.
4-3 are drawn to satisfy this requirement. Certain vibrations in a molecule are
referred to as normal vibrations or normal modes. These are independent,
self-repeating displacements of the atoms that preserve the center of mass. In a
normal vibration, all the atoms vibrate in phase and with the same frequency.
It is possible to resolve the most complex molecular vibration into a relatively
4-3 Atomic and Molecular Transitions 93
t FIGURE 4-3 Two different

S S S S vibrations for the SO2
0 0 O molecule. (The amplitudes
and angle deformations are
0 0 exaggerated to illustrate the
mode.)
(A) (B)
small number of such normal modes. For a non-linear molecule, there are 3N - 6
such modes, where N is the number of atoms in the molecule. In Chapter 6, the
procedure for calculating the total number of normal modes and their symmetries
will be given. The normal modes can be considered as the 3N - 6 internal degrees
of freedom that (in the absence of anharmonicity) could take up energy
independently of each other. The motion of the atoms of a molecule in the different
normal modes can be described by a set of normal coordinates. These are a set
of coordinates defined so as to describe the normal vibration most simply. They
often are complicated functions of angles and distances.
The products of the normal modes of vibration are related to the total
vibrational state, 0, of a non-linear molecule as shown in equation (4-5).
3N -6
Or= H n=1
On (4-5)
where LI indicates that the product of the n vibrational modes is to be taken and
0. is the wave function for a given normal mode. There exists for each normal
vibration (On) a whole series of excited vibrational states, i, whose harmonic
oscillator wave functions are given by:
= Ni exp[(- 1)ajq 2 ]Hj( aq) (4-6)
where i = 0, 1, 2 ... ; Hi is the Hermite polynomial of degree i; aj = 2nrvi/h;

N= [ i!
1
)] 1 2; and q is the normal coordinate.
7(2
From this, the following wave functions are obtained for the ground, first,
and second excited states:
0 (o)'1 exp[( -)aq2] (4-7)
i = 2() q exp[(- ')aiq (4-8)
0 2 = 2(a2) (2a 2 q 2 - 1)exp(- j)a 2 q2] (4-9)
When 0, is plotted as a function of displacements about the normal

coordinate, q, with zero taken as the equilibrium internuclear distance, Fig. 4-4
is obtained. This symmetric function results because q appears only as q2 . The
same type of plot is obtained for the ground state of all normal modes. It is
totally symmetric; accordingly, the total vibrational ground state of a molecule
%PO
must belong to the totally symmetric irreducible representation, for it is the
product of only totally symmetric vibrational wave functions.
0 The vibiational excited state function 01 has a functional dependence on q
q
that is not of an even power, and accordingly it does not necessarily have a,
symmetry. When one normal mode is excited in a vibrational transition, the
FIGURE 4-4 A plot of the resulting total state is a product of all the other totally symmetric wave functions
ground vibrational wave
and the wave function for this first vibrational excited state for the excited normal
function versus the normal
coordinate.
mode. Thus, the total vibrational state has symmetry corresponding to the
normal mode that is excited.
The rotationalstates correspond to quantized molecular rotation around an
axis without any appreciable change in bond lengths or angles. Different rotational
states correspond to different angular momenta of rotation or to rotations about
different axes. Rotation about the C 2 axis in SO 2 is an example of rotational
motion.
In the treatment of molecular spectra, the Born-Oppenheimer approximation
is invoked. This approximation proposes that the total energy of a system may
be regarded as the sum of three independent energies: electronic, vibrational, and
rotational. For example, the electronic energy of the system does not change as
vibration of the nuclei occurs. The wave function for a given molecular state can
then be described by the product of three independent wave functions: 0., 0,ib,
and 0. As we shall see later, this approximation is not completely valid.
The relative energies of these different molecular energy states in typical
molecules are represented in Fig. 4-5. Rotational energy states are more closely
spaced than are vibrational states, which, in turn, have smaller energy differences
than electronic states. The letters v., vi, etc., and v,', v,', etc., each represent the
vibrational levels of one vibrational mode in the ground and first electronic
excited states, respectively. Ultraviolet or visible radiation is commonly required
to excite the molecule into the excited electronic states. Lower-energy infrared
suffices for vibrational transitions,while pure rotational transitionsare observed in
the still lower-energy microwave and radio-frequency regions.
Electronic transitions are usually accompanied by changes in vibration and
rotation. Two such transitions are indicated by arrows (a) and (b) of Fig. 4-5.
In the vibrational spectrum, transitions to different rotational levels also occur.
As a result, vibrational fine structure is often detected in electronic transitions.
V,
1st excited electronic state
FIGURE 4-5 Energy states of

a diatomic molecule. (a) (b)
V3
2 Ground electronic state

v1
V0
4-5 Relaxation and Chemical Exchange Influences on Spectral Line Width 95
Rotational fine structure in electronic transitions can be detected in high

resolution work in the gas phase. Rotational fine structure in vibrational
transitions is sometimes observed in the liquid state and generally in the gaseous
state.
The energy level diagram in Fig. 4-5 is that for a diatomic molecule. For a
polyatomic molecule, the individual observed transitions can often be described
by diagrams of this type, each transition in effect being described by a different
diagram.
4-4 SELECTION RULES
In order for matter to absorb the electric field component of radiation, another
requirement in addition to energy matching must be met. The energy transition
in the molecule must be accompanied by a change in the electrical center of the
molecule in order that electrical work can be done on the molecule by the
electromagnetic radiation field. Only if this condition is satisfied can absorption
occur. Requirements for the absorption of light by matter are summarized in the
selection rules. Transitions that are possible according to these rules are referred
to as allowed transitions, and those not possible asforbidden transitions. It should
be noted, however, that the term "forbidden" refers to rules set up for a simple
model and, while the model is a good one, "forbidden" transitions may occur
by mechanisms not included in the simple model. The intensity of absorption or
emission accompanying a transition is related to the probability of the transition,
the more probable transitions giving rise to more intense absorption. Forbidden
transitions have low probability and give absorptions of very low intensity. This
topic and the symmetry aspects of this topic will be treated in detail when we
discuss the various spectroscopic methods.
4-5 RELAXATION AND CHEMICAL EXCHANGE

INFLUENCES ON SPECTRAL LINE WIDTH
Spectral transitions involve different atomic or molecular energy states. The

Boltzmann expression indicates the probability, Pi, that any molecule is in the
ith state as a function of the ground state energy E0 and the excited state energy E,.
e- (E,- E)/kT
Pi =
-e (Ei - E)|kT
Here, k is the Boltzmann constant (1.3807 x 10-" JK-') and T the absolute
temperature.
When the energy differences of the states are large as they are in most
electronic and vibrational transitions, only the ground state is populated at room
temperature. When a photon is absorbed, the molecule changes to the excited
state and then relaxes back to its ground state by a radiative process that emits /
radiation or increases the kinetic energy (temperature) of the sample. In time, the
Boltzmann equilibrium is reestablished. When the energy difference of the states
is small compared to kT at room temperature (as is the case for states involved
in an nmr transition), the states are populated in accord with the Boltzmann
expression. A dynamic system results in which molecules in the excited state
exchange energy with other identical molecules. The fraction of molecules in the
two states is constant, but a given molecule exchanges between the different states.
of
rThe process of switching energy states is called relaxation and the efficiency
I interchange is governed by the relaxation mechanism, which in turn governs the
lifetime of the molecule in a given state. When a molecule is excited by the
absorption of energy, these relaxation mechanisms determine the lifetime of the
excited state as the system returns to the Boltzmann equilibrium. The natural
fine width of a spectral line depends upon the lifetime of the excited state via the
uncertainty principle.
AEAt ~ h
or
1
AvAt ~- (4-10)
27r
As the lifetime, At, decreases, the range of frequencies possible, i.e., the natural
line width, increases.
Next, consider the consequences of chemical reactions on the widths of
spectral lines. When two or more chemically distinct species coexist in rapid
equilibrium (two conformations, or rapid chemical exchange, e.g., proton exchange
between NH 3 and NH 4 , or other equilibria), one often sees absorption peaks
corresponding to the individual species in some forms of spectroscopy; but with
other methods, only a single average peak is detected. A process may cause a
broadening of the spectral line in some spectral regions, while the same process
in some other spectral region has no effect.
Consider the case in which two different sites give rise to two distinct peaks.
As the rate of the chemical exchange comes close to the frequency Av of the
spectroscopic method, the two peaks begin to broaden. As the rate becomes
faster, they move together, merge, and then sharpen into a single peak. In
subsequent chapters, we shall discuss procedures for extracting rate data over
this entire range. Here we shall attempt to gain an appreciation for the time
scales corresponding to the various methods by examining the rates that result
in the line broadening of a sharp single resonance and the merging of two distinct
resonances.
First, we shall be concerned with the broadening of a resonance line of an
individual species. If At is the lifetime of the excited state, we have:
1
Av(sec -) ~ 1
2nrAt(sec)
If there is some chemical or physical process that is fast compared to the excited
state lifetime, At can be shortened by this process and the line will be broadened.
4-5 Relaxation and Chemical Exchange Influences on Spectral Line Width 97
In infrared, for example, it is possible to resolve* two bands corresponding to
different sites that are separated by 0.1 cm -1, so we can substitute this value into
equation (4-10) and find that lifetimes, At, corresponding to this are given by:
Av = 0.1 cm-' x 3 x 1010 cm/sec = 3 x 10' sec-1
1
At -5 x 10" sec
(2)(n)(3 x 10' sec- ' 5)
Therefore, we need a process that will give rise to lifetimes of - 10 sec or less
to cause broadening. Since lifetimes are the reciprocals of first order rate constants,
this means we need a process whose rate constant is at least 101" sec-1 in order
to detect a broadening in the infrared band. Diffusion-controlled chemical
reactions have rate constants of only 1010, so for systems undergoing chemical
exchange, the chemical process will have no influence on the infrared line shape.
Rotational motion occurs on a time scale that causes broadening of an infrared
or Raman line, and it is possible to obtain information about the motion from
the line shape.m) The inversion doubling of ammonia is also fast enough to affect
the Raman and microwave spectra.
In nmr, typical resolution is ~0.1 Hz (cycle per second), which, when
substituted into the above equation, shows that a process with a lifetime of about
2 sec (or less) or a rate constant of about 0.5 sec 1 (or more) is needed to broaden
the spectral line. This is in the range of many chemical exchange reactions.
Next, we shall consider the rate at which processes must occur to result in
a spectrum in which only an average line is detected for two species. In order
for this to occur, the species must be interconverting so fast that the lifetime in
either one of the states is less than At, so that only an average line results. This
is calculated by substituting the difference between the frequencies of the two
states for Av in equation (4-10). In the infrared, a rate constant of 5 x 10' " sec-
or greater would be required to merge two peaks that are separated by 300 cm
1 1
At 2 0- e
(2)(n)(9.0 x 1012 sec -)- 2 x 10 se
In nmr, for a system in which, for example, the two proton peaks are separated
by 100 Hz, one obtains
1
At 2 x 10- sec
(2)(7r)(100 sec 2)
Thus, an exchange process involving this proton that had a rate constant of
5 x 102 sec-1 would cause these two resonances to appear as one broad line.
As the rate constant becomes much larger than 5 x 102 (e.g., by raising the
temperature for a process with a positive activation enthalpy), the broadened
* By resolve, we mean that one can detect the existence of two distinct maxima. We are using
this quantity to roughly estimate Av, the time scale of the spectroscopic method (i.e., the natural line
width).
TABLE 4-1 Kinetic Techniques and the Appropriate Lifetimes*
conventional kinetics 10 sec or longer (determined by ability to

mix reagents)
stop flow 10'secorlonger(fastermixingispossible)
nmr 10 to 10- sec
esr 10 to 1 sec
temperature jump 0.1 to 10-6 sec
M6ssbauer (iron) 10 - sec
ultrasonics 10-4 to 10-8 sec
fluorescent polarization (the depolarization 10-8 to 10-' sec
is measured)
IR and Raman line shapes 10-" sec
photoelectron spectroscopy 10 -18 Sec
* If some of these techniques are unfamiliar at present, do not be concerned. This will
provide a ready reference after you encounter them in your study.
single resonance begins to sharpen. Eventually, as the process becomes very fast,
the line width is no longer influenced by the chemical process. The radiation is
too slow to detect any chemical changes that are occurring and a sharp average
line results. This is analogous to the eye being too slow to detect the electronic
sweep that produces a television picture.
In x-ray diffraction, the frequency of the radiation
to" is sec'. Since this
is too fast, compared with molecular rearrangements, all we would detect for a
dynamic system would be disorder, i.e., contributions from each of the dominant
conformations.
In M6ssbauer spectroscopy, we observe a very high energy process in which
a nucleus in the sample absorbs a y-ray from the source. The lifetime of the
nuclear excited state, for iron is constant at 10- sec. Thus, if there is a chemical
process occurring in an iron sample that equilibrates two iron atoms with a first
order rate constant greater than 107 sec1 , the Mossbauer spectrum will reveal
only an average peak. In dichlorobisphenanthroline iron (111), the equilibrium
mixture of high spin (five unpaired electrons) and low spin (one unpaired electron)
iron complexes interconvert so rapidly (they have a lifetime < 10- sec) that only
a single iron species is observed in the spectrum. For a ruthenium nucleus, the
lifetime of the nuclear excited state is 10 a sec; this determines the time scale for
Applicattions experiments with this nucleus. A similar situation pertains in x-ray photoelectron
spectroscopy. The time scale of an electron is not influenced by chemical processes
and the lifetime for the resulting excited state is constant at about 10 18 sec.
Table 4-1 summarizes the methods used in kinetic studies and the ranges
of lifetimes that can be investigated.
General The following general applications of spectroscopy are elementary and pertain
Stoboth vibrational and electronic spectroscopic methods: (1) determination of
concentration, (2) "fingerprinting," and (3) determination of the number of species
in solution by the use of isosbestic points.
4-6 DETERMINATION OF CONCENTRATION
Measurement of the concentrations of species has several important applications.

If the system is measured at equilibrium, equilibrium constants can be determined.
By evaluating the equilibrium constant, K, at several temperatures, the enthalpy
AH' for the equilibrium reaction can be calculated from the van't Hoff equation:
- AH 0
log K = - + C (4-11)
2.3RT
Determination of the change in concentration of materials with time is the basis

of kinetic studies that give information about reaction mechanisms. In view of
the contribution of results from equilibrium and kinetic studies to our under-
standing of chemical reactivity, the determination of concentrations by spectro-
scopic methods will be discussed.
The relationship between the amount of light absorbed by certain systems
and the concentration of the absorbing species is expressed by the Beer-Lambert
law:
A = logi 0 = rcb (4-12)
where A is the absorbance, I is the intensity of the incident light, I is the intensity
of the transmitted light, e is the molar absorptivity (sometimes called extinction
coefficient) at a given wavelength and temperature, c is the concentration (the
molarity if Eis the molar absorptivity), and b is the length of the absorbing system.
The molar absorptivity varies with both wavelength and temperature, so these
must be held constant when using equation (4-12). When matched cells are
employed to eliminate scattering of the incident beam, there are no exceptions
to the relationship between absorbance and b (the Lambert law). For a given
concentration of a certain substance, the absorbance is always directly propor-
tional to the length of the cell. The part of equation (4-12) relating absorbance
and concentration (log 10 I/I = ec, for a constant cell length) is referred to as
Beer's law. Many systems have been found that do not obey Beer's law. The
anomalies can be attributed to changes in the composition of the system with
concentration (e.g., different degrees of ionization or dissociation of a solute at
different concentrations). For all systems, the Beer's law relationship must be
demonstrated, rather than assumed, over the entire concentration range to be
considered. If Beer's law is obeyed, it becomes a simple matter to use equation
(4-12) for the determination of the concentration of a known substance if it is
the only material present that is absorbing in a particular region of the spectrum.
The Beer's law relationship is tested and e determined by measuring the
absorbances of several solutions of different known concentrations covering the
range to be considered. For each solution, E can be calculated from
A(1 cm cell) (4-13)

c, molarity
where the units of e are liters mole 1 cm -. For a solution of this material of
unknown concentration, the absorbance is measured, Eis known (4-13), and c is
calculated from equation (4-12).
There are some interesting variations on the application of Beer's law. If the
absorption of two species should overlap, this overlap can be resolved math-
ematically and the concentrations determined. This is possible as long as the two
e values are not identical at all wavelengths. Consider the case in which the r
values for two compounds whose spectra overlap can be measured for the pure
compounds. The concentration of each component in a mixture of the two
compounds can be obtained by measuring the absorbance at two different
wavelengths, one at which both compounds absorb strongly and a second at
which there is a large difference in the absorptions. Both wavelengths should be
selected at reasonably flat regions of the absorption curves of the pure compounds,
if possible. Consider two such species B and C, and wavelengths ), and A2 . There
are molar absorptivities of rBe for B at A1, eBA2 for B at A2 , and similar quantities
EC and ECA for C. The total absorbance of the mixture at A1 is A,, and that at
A2 is A2 . It follows that:
A1 = XEBA + YE;](4-14)
and
A 2 = XEBA2 + 2 (4-15)
where x = molarity of B and y = molarity of C. The two simultaneous equations

(4-14) and (4-15) have only two unknowns, x and y, and can be solved.
A situation often encountered in practice involves an equilibrium of the type
D + E ;r2 DE
Here the spectra of, say, D and DE overlap but E does not absorb; if the
equilibrium constant is small, pure DE cannot be obtained, so its molar
absorptivity cannot be determined directly. It is possible to solve this problem
for the equilibrium concentrations of all species, i.e., to determine the equilibrium
constant:
K =K-[DE]
[D](4-16)
[D][E]
Let [D], be the initial concentration of D. This can be measured, as can [E]0 ,
the initial concentration of E. The molar absorptivity of DE, VDE, cannot be
determined directly, but it is assumed that Beer's law is obeyed. It can be seen
from a material balance that:
[D] = [D]O - [DE] (4-17)
[E] = [E]O - [DE] (4-18)

so
K
K =DE [DE] (4-19)
[D]o - [DE])([E], - [DE])
The total absorbance consists of contributions from [D] and [DE] according to
A = ED[D] + EDE[DE] (4-20)
Substituting equation (4-17) into (4-20) and solving for [DE], one obtains
A - ED[D]o
[DE] =
EDE ED
but eD[D]o is the initial absorbance, Ao, of a solution of D with concentration

[D], without any E in it, so
A - A0
[DE] - (4-21)
EDE -D
When equation (4-21) is substituted into the equilibrium constant expression,

equation (4-19), and this is rearranged, we have 2 )
K-1 = - [D31 - [Elo + [D]O[E]o(EDE - D)

EDE -D A - A
The advantage of this equation is that it contains only two unknown quantities,
EDE and K-'. Furthermore, these unknowns are constant for any solution of
different concentrations [D]o and [E]O that we care to make up. For two different
sets of experimental conditions (different values of [D], and [E]O) two simulta-
neous equations can be solved; CDE is eliminated, and K is obtained. If several
sets of experimental conditions are employed, all possible combinations of
simultaneous equations can be considered by employing a least-squares computer
analysis"-" that finds the best values of K-1 and EDE to reproduce the
experimental absorbances. These least-squares procedures usually provide an
error analysis that is most important to have (vide infra).
There are several advantages to a graphical display of the simultaneous
equations that are being analyzed. The graph is constructed(') by taking a solution
of known [D],, [E],, and A, and calculating the values of K - that would result
1
by selecting several different values of CDE near the expected value. These results
are plotted, as in Fig. 4-6, by the line [D][E],*. The procedure is repeated for
other initial concentrations, e.g., [D]'[E],' and [D]o"[E]o". The intersection of
any two of these lines is a graphical representation of the solution of two
simultaneous equations. The intersection of all the calculated curves should occur
at a point whose values of K and E satisfy all the experimental data. This
common intersection justifies the Beer's law assumption used in the derivation
because it indicates a unique value for E for all concentrations. As a result of
experimental error, a triangle usually results instead of a point. The best K-
and Eto fit the data are then found by a least-squares procedure. References 6 and
[D]6'[E]"'
[D]06[E]6
[DIO[Elo
FIGURE 4-6 Graphical solution of K-1

equation (4-22).
7 describe the least-squares procedure and error analysis. When the experiment
described above is not designed properly, a set of concentrations of D and E are
employed that result in a set of parallel lines for the K-1 vs. e plots. The slope
of one of these lines is given by taking the partial derivative of the K 1 expression
in (4-22) with respect to EDE - ED:
OK A - A0 [D]o[E]o
= - _ + (4-23)
BEDE - D (VDE )2 A - A0
Since the first term is generally small, we see from equation (4-23) that if
experimental conditions are selected for a series of experiments in which [DE]
or the value of A nearly doubles every time [D], or [E], is doubled, then the
various resulting K-1 vs. r plots will be nearly parallel. The values of K-1 and
r obtained from the computer analysis of this kind of data(6) are thus highly
correlated and should be considered undefined, for the simultaneous equations
are essentially dependent ones. The K - vs. r plots have been described in terms
of the results from the error analysis of the least-squares calculation, and the
reader is referred to reference 7 for details.
The approach described above for the evaluation of equilibrium constants
is a general one that applies to any form of spectroscopy or any kind of
measurement in which the measured quantity is linearly related to concentration
(i.e., in which a counterpart to equation (4-20) exists). Similar analyses have been
7
described for calorimetric data(8) and for nmr spectral data. )
There have been many reports in the literature of attempts to solve
simultaneous equations for equilibrium constants for two or more consecutive
equilibria:
A + B:#AB step I
AB + Be AB 2 step 2
Usually 8AB, FAB 2 , and the stepwise equilibrium constants K 1 and K 2 are unknown.
In most instances, even though the reported parameters fit the experimental data
well, the system is undefined. Careful examination often shows that many other,
very different combinations of parameters also fit the data. To solve this problem,
one must find a region of the spectrum in which AB makes the main contribution
to the absorbance and another in which AB 2 makes the predominant contribution.
Only by working at both wavelengths can one solve for all the unknowns in a
rigorous fashion. This is impossible in many nmr, u.v.-visible, and calorimetric
experiments because properties related predominantly to the individual AB and
AB 2 species cannot be monitored separately. This problem has been discussed
in detail in the literature"'1 and several examples are given there.
4-7 ISOSBESTIC POINTS
If two substances at equal concentrations have absorption bands that overlap,

there will be some wavelength at which the molar absorptivities of the two species
are equal. If the sum of the concentrations of these two materials in solution is
held constant, there will be no change in absorbance at this wavelength as the
ratio of the two materials is varied. For example, assume a reaction A + B -* AB,
in which only A and AB absorb; the sum of [A] plus [AB] will be constant as
long as the initial concentration of A is held constant as B is varied. Since the
absorbance of the solution is given by
abs = EA[A] + A[AB] (4-24)
the absorbance will be constant when 8A =AB and [A] plus [AB] is held constant.
The invariant point obtained for this system is referred to as the isosbestic point.
The existence of one or more isosbestic points in a system provides information
regarding the number of species present. For example, the spectra in Fig. 4- 7
were obtained( 9 ) by keeping the total iron concentration constant in a system
consisting of:
Curve 1 2.1 moles of LiCl per Fe(DMA)(Cl0 4 )3
Curve 2 2.6 moles of LiCI per Fe(DMA)(Cl0 4 ) 3
Curve 3 3.1 moles of LiCl per Fe(DMA),(ClO 4 )3
Curve 4 4.1 moles of LiCl per Fe(DMA),(ClO 4 ),
where DMA is the abbreviation for N,N-dimethylacetamide. Points A and B are
isosbestic points; their existence suggests that the absorption in this region is
essentially accounted for by two species. Curve 4 is characteristic of FeCl 4 -. A
study of solutions more dilute in chloride establishes the existence of a species
Fe(DMA)4 Cl 2 +, which exists at a 2:1 ratio of Cl- to Fe." and absorbs in this
region. The isosbestic points indicate that the system can be described by the
species Fe(DMA)4 Cl2 + and FeCl4 - over the region from 2:1 to 4: 1 ratios of
Cl to Fe'". The addition compound FeCl3 -DMA probably does not exist in
appreciable concentration in this system, i.e., at total Fe"' concentrations of
2 x 10- 4 M. Curve 4 in this spectrum does not pass through the isosbestic points.
Small deviations (e.g., curve 4 at point A) may be due to experimental inaccuracy,
changes in solvent properties in different solutions, or a small concentration of
a third species, probably FeCl3 -DMA, present in all systems except that
represented by curve 4.
The conclusion that only two species are present in appreciable concentra-
tions could be in error if FeC13 -DMA or other species were present that had
molar absorptivities identical to that of the above two ions at the isosbestic
points. However, if equilibrium constants for the equilibrium
Fe(DMA)4 Cl2 + + 2Cl- FeCI4 - + 4DMA

FIGURE 4-7 Spectra of the

Fe(DMA) 4Cl2+-LiCI system in
N,N-dimethylacetamide as
solvent.
X (mp)
are calculated from these different curves at different wavelengths, the possibility
of a third species is eliminated if the constants agree.
An interesting situation results in which an isosbestic point can be obtained
in solution when more than two species with differing extinction coefficients exist
if the base (or acid) employed has two donor (or acceptor) sites. For example, if
DMA were to coordinate to an acid A to produce an oxygen-bound complex
and a nitrogen-bound one, the mixture of complexes AN (nitrogen bound), AO
(oxygen bound), and free A (an absorbing Lewis acid) will give rise to an isosbestic
point."') The absorbance for such a system is given by equation (4-25):
abs = CAA] + Ao[AO] + rA[AN] (4-25)
The equilibrium constant expressions are given by
[AO] KN =
[AN]
[N
and
0 [A][B] N [A][B]
[AO] + [A N] [AB]
N [A][B] (4-26)
[A][B]
where we define [AO] + [AN] = [AB]. The fraction of complex that is oxygen-
coordinated, XAO, is given by
[AO]
XAO - Ko [A][B] [AO] (4-27)
KO + KN [AB] [AB]
[A][B]
The fraction that is nitrogen-coordinated is similarly derived as:
XAN [AN] (4-28)

[AB]
Now the total absorbance, abs, becomes
abs= 4[A] + EAOXAO[AB] + VANXAN[AB]
or
abs = F[A] + ('AOXAO + £ANXAN)[AB] =A[A] + t{'[AB] (4-29)
Since the sum of [A] and [AB] is constant, and there must be a point in the
overlapping spectra at which rs=', an isosbestic point will be obtained even
though three absorbing species exist. This is true because
XAN
- -
KN
- a constant
XAo Ko
We have taken two absorbing species whose ratio is a constant, independent

of the parameter being varied, and we have translated them into what is effectively
a single absorbing species via equation (4-29). The general criterion is thus that
2 + N absorbing species will give rise to an isosbestic point if there are N
independent equations of the form
[Y]= k
(4-30)
[Z]
where Y and Z are two of the absorbing species in the system and the value of
k is independent of the parameter being varied.
As an example of the application of this criterion, consider an acid that can
form two isomers with a base, AB and A*B (e.g., Cu(hfac) 2 forming basal and
apical adducts); let this acid form an adduct with a base that can form two
isomers with an acid, AB and AB* (e.g., N-methyl imidazole bound through the
amine and imine nitrogens). Five absorbing species (A, AB, A*B, AB*, and A*B*)
can exist,
A + B -- AB
A + B -- A*B
A + B -- AB*
A + B - A*B*
so three constants are needed. The equilibrium constants are
[AB] [A*B] [AB*1 [A*B*]

[A][B] 2 [A][B] 3 [A][B] [A][B]
We now note that three independent ratios exist, which are independent of [B]:
K1 [AB] K, [A*B] K 3 [AB*]

K2 [A*B] K3 [AB*] K4 [A*B*]
Therefore, an isosbestic point is expected if [A], is held constant and [B], is

varied. Any other ratio of equilibrium constants, e.g., K 1/K,, is not independent
of the three ratios written. The general rules presented here apply to a large
number of systems. However, rote application of rules is no substitute for an
understanding of the systems under consideration.
4-8 JOB'S METHOD OF ISOMOLAR SOLUTIONS
By examining the spectra of a series of solutions of widely varying mole ratios

of A to B, but with the same total number of moles of A + B, the stoichiometry
of complexes formed between A and B in solution can often be determined.
Absorbance at a wavelength of maximum change is plotted against the mole
ratio of A to B; the latter is usually used as the abscissa. The plots have at least
one extremum, often a maximum. In simple cases, the extrema occur at mole
fractions corresponding to the stoichiometry of the complexes that form in
solution. 1-13)
4-9 "FINGERPRINTING"
This technique is useful for the identification of an unknown compound that is

suspected to be the same as a known compound. The spectra are compared with
respect to r values, wavelengths of maximum absorption, and band shapes.
In addition to this direct comparison, certain functional groups have
characteristic absorptions in various regions of the spectrum. For example, the
carbonyl group will generally absorb at certain wavelengths in the ultraviolet
and infrared spectra. Its presence in an unknown compound can be determined
from these absorptions. Often one can even determine whether or not the carbonyl
group is in a conjugated system. These details will be considered later when the
spectroscopic methods are discussed individually.
Spectroscopic methods provide a convenient way of detecting certain
impurities in a sample. For example, the presence of water in a system can easily
be detected by its characteristic infrared absorption. Similarly, a product can be
tested for absence of starting material if the starting material has a functional
group with a characteristic absorption that disappears during the reaction.
Spectral procedures are speedier and less costly than elemental analyses. The
presence of contaminants in small amounts can be detected if their molar
absorptivities are large enough.
Exercises 107
1. E. F. Johnson and R. S. Drago, J. Amer. Chem. Soc., 95, 1391 (1973), and references REFERENCES
therein.
2. N. J. Rose and R. S. Drago, J. Amer. Chem. Soc., 81, 6138 (1959). CITED
3. T. D. Epley and R. S. Drago, J.Amer. Chem. Soc., 91,2883 (1969), and references therein.
4. T. 0. Maier and R. S. Drago, Inorg. Chem., 11, 1861 (1972).
5. P. J. Lingane and Z. Z. Hugus, Jr., Inorg. Chem., 9, 757 (1970).
6. R. M. Guidry and R. S. Drago, J. Amer. Chem. Soc., 95, 6645 (1973).
7. F. L. Slejko, R. S. Drago, and D. G. Brown, J. Amer. Chem. Soc., 94, 9210 (1972).
8. T. F. Bolles and R. S. Drago, J. Amer. Chem. Soc., 87, 5015 (1965)..
9. R. L. Carlson, K. F. Purcell, and R. S. Drago, Inorg. Chem., 4, 15 (1965).
10. R. G. Mayer and R. S. Drago, Inorg. Chem., 15, 2010 (1976).
11. W. Likussar and D. F. Boltz, Anal. Chem., 43, 1265 (1971).
12. M. M. Jones and K. K. Innes, J. Phys. Chem., 62, 1005 (1958).
13. E. Asmus, Z. Anal. Chem., 183, 321, 401 (1961).
1. A series of different molecular transitions require the following energies in order to EXERCISES
occur. Indicate the spectral region in which you would expect absorption of radiation
to occur, and the wavelength of the radiation
a. 0.001 kcal mole-1.
b. 100 kcal mole 1.
c. 30 kcal mole - 1.
2. Convert the wavelength units in part a of exercise I to wavenumbers (cm-1).
3. Convert the following wavenumbers to p and A:

a. 3600 cm
b. 1200 cm
4. Convert 800 mp to cm-'.
5. The r value for compound X is 9000 liters mole 1 cm'. A 0.1 molar solution of X in
water, measured in a 1 cm cell, has an absorbance of 0.542. It is known that X reacts
according to the equation:
X±Y + Z
X obeys Beer's law, and Y and Z do not absorb in this region. What is the equilibrium
constant for this reaction?
6. For the equilibrium
A + B:C
assume that only C absorbs and that EC cannot be measured. Derive an equation similar
to (4-22) for this system.
7. a. Use equations (4-14) and(4-15)and values of Al = 0.3, A 2 = 0.7, es = 5, 2 = 12,

8c;' = 13, and ee, = 2 to construct a plot similar to that in Fig. 4-6 to illustrate
the graphical solution of the two simultaneous equations.
b. Construct the same plot as in part a, using A1 = 0.3, A 2 = 0.7, B 5, -,

e = 12,
c = 3, and Fc = 5. Compare your confidence in the two answers, realizing that
there are 2 to 30% error in the A and e values.
c. Equation (4-22) is somewhat analogous to that used in parts a and b, for the first
term is generally small under the conditions used in a spectroscopic study. Often
investigators work in a concentration range in which, holding [D], constant and
doubling [E]O, they double the amount of complex formed or they double A - A0 .
What would the two lines in the K-1 vs. e plot look like?
8. A proton is being rapidly exchanged between B and B', i.e.,
BH + + B'± B'H + + B
The frequency difference between the H * peaks of BH + and B'H + in the nmr spectrum is
100 Hz (cycles per second).
a. What rate of exchange would be required for the two peaks to appear as a single peak?
b. Would a slower or faster rate be required to observe two separate resonances that are
broadened by the exchange process?
c. What is the approximate maximum rate of chemical exchange that could occur and
still have no effect on the nmr spectrum?
9. A system A + B: C + D is investigated spectroscopically. Species A, C, and D all

have absorbances that overlap. Since there are more than two absorbing species, would
one expect to find an isosbestic point if this were the only equilibrium involved? Prove
your answer.
Electronic Absorption
Spectroscopy
5-1 VIBRATIONAL AND ELECTRONIC ENERGY LEVELS Introduction

IN A DIATOMIC MOLECULE
Prior to a discussion of electronic absorption spectroscopy, the information

summarized by a potential energy curve for a diatomic molecule (indicated in
Fig. 5-1) will be reviewed. Fig. 5-1 is a plot of E, the total energy of the system,
versus r, the internuclear distance, and is one of many types of potential functions
referred to as a Morse potential. The curve is expressed mathematically by
V = D{1 - exp[- v(2nr2 piD)' 2 (r - r,)]}2
All terms are defined in Fig. 5-1 except p, which is the reduced mass,
(m1m 2)(mi 1 + m2)-
As the bond distance is varied in a given vibrational state, e.g., along A-B
in the v 2 level, the molecule is in a constant vibrational energy level. At A and
B we have, respectively, the minimum and maximum values for the bond distance
in this vibrational level. At these points the atoms are changing direction, so the
vibrational kinetic energy is zero and the total vibrational energy of the system
is potential. At re, the equilibrium internuclear distance, the vibrational kinetic
Continuum
BDC/FIGURE 5-1 Morse energy
uw A v2 Dissociation curve for a diatomic molecule.
v, Energy, D
Zero { vO
point
energy
0 r,
Internuclear distance, r 109
110 Chapter 5 Electronic Absorption Spectroscopy
energy is a maximum and the vibrational potential energy is zero. Each horizontal
line represents a different vibrational energy state. The ground state is vo and
excited states are represented as vI, v 2 , etc. Eventually, if enough energy can be
absorbed in vibrational modes, the molecule is excited into the continuum and
it dissociates. For most compounds, nearly all the molecules are in the vo level
at room temperature because the energy difference v, - vo is usually much larger
than kT (thermal energy), which has a value of 200 cm- at 300 K. (Recall the
Boltzmann expression.)
Each excited electronic state also contains a series of different vibrational
energy levels and may be represented by a potential energy curve. The ground
electronic state and one of the many excited electronic states for a typical diatomic
molecule are illustrated in Fig. 5-2. Each vibrational level, v., is described by a
vibrational wave function, fib. For simplicity only four levels are indicated. The
square of a wave function gives the probability distribution, and in this case iV1b2
indicates probable internuclear distances for a particular vibrational state. This
function, Ivib2, is indicated for the various levels by the dotted lines in Fig. 5-2.
The dotted line is not related to the energy axis. The higher this line, the more
probable the corresponding internuclear distance. The most probable distance
for a molecule in the ground state is re, while there are two most probable
distances corresponding to the two maxima in the next vibrational energy level
of the ground electronic state, three in the third, etc. In the excited vibrational
levels of the ground and excited electronic states, there is high probability of the
molecule having an internuclear distance at the ends of the potential function.
FIGURE 5-2 Morse curves

for a ground and excited
state of a diatomic molecule, C
showing vibrational
2
probability function, //ib , as
dotted lines.
re
internuclear distance, r
5-2 Relationship of Potential Energy Curves to Electronic Spectra 111
a,
a 0--2 FIGURE 5-3 Spectrum
corresponding to the potential
energy curves indicated in Fig.
5-2.
Wavelength, X
5-2 RELATIONSHIP OF POTENTIAL ENERGY CURVES TO

ELECTRONIC SPECTRA
An understanding of electronic absorption spectroscopy requires consideration

of three additional principles:
1. In the very short time required for an electronic transition to take place
(about 10- " sec), the atoms in a molecule do not have time to change position
appreciably. This statement is referred to as the Franck-Condonprinciple.* Since
the electronic transition is rapid, the molecule will find itself with the same
molecular configuration and vibrational kinetic energy in the excited state that
it had in the ground state at the moment of absorption of the photon. As a result,
all electronic transitions are indicated by a vertical line on the Morse potential
energy diagram of the ground and excited states (see the arrows in Fig. 5-2); i.e.,
there is no change in internuclear distance during the transition.
2. There is no general selection rule that restricts the change in vibrational
state accompanying an electronic transition. Frequently transitions occur from
the ground vibrational level of the ground electronic state to many different
vibrational levels of a particular excited electronic state. Such transitions may
give rise to vibrational fine structure in the broader peak of the electronic
transition.
The three transitions indicated by arrows in Fig. 5-2 could give rise to three
peaks. Since nearly all of the molecules are present in the ground vibrational
level, nearly all transitions that give rise to a peak in the absorption spectrum
will arise from vo. Transitions from this ground level (vo) to v,', v,', or v2 ' are
1
referred to as 0 -* 0, 0 -- 1, or 0 -- 2 transitions, respectively. It can be shown )
that the relative intensity of the various vibrational sub-bands depends upon the
vibrational wave function for the various levels. A transition is favored if the
* This principle was originally proposed by Franck. It was given a quantum mechanical
interpretation and extended by Condon.
tIn the gas phase, various rotational levels in the ground vibrational state will be populated
and transitions to various rotational levels in the excited state will occur, giving rise to fine structure
in the spectrum. This fine structure is absent in solution because collision of the solute with a solvent
molecule occurs before a rotation is completed. Rotational fine structure will not be discussed.
probabilities of the ground and excited states of the molecule are both large for
the same internuclear distance. Three such transitions are indicated by arrows
in Fig. 5-2. The spectrum in Fig. 5-3 could result from a substance in solution
undergoing the three transitions indicated in Fig. 5-2. The 0 -* 0 transition is the
lowest energy-longest wavelength transition. The differences in wavelength at
which the peaks occur represent the energy differences of the vibrational levels
in the excited state of the molecule. Much information about the structure and
configuration of the excited state can be obtained from the fine structure.
Electronic transitions from bonding to antibonding molecular orbitals are
often encountered. In this case the potential energy curve for the ground state
will be quite different from that of the excited state because there is less bonding
electrondensity in the excited state. As a result, the equilibrium internuclear
distance will be greater and the potential energy curve will be broader for the
excited state. Because of this displacement of the excited state potential energy
curve, the 0 -, 0 and transitions to other low vibrational levels may not be
observed. A transition to a higher vibrational level becomes more probable. This
can be visualized by broadening and displacing the excited state represented in
Fig. 5-2.
3. There is an additional symmetry requirement, which was neglected for
the sake of simplicity in the above discussion. It has been assumed that this
symmetry requirement is satisfied for the transitions involved in this discussion.
This will be discussed subsequently in the section on selection rules.
The above discussion pertains to a diatomic molecule, but the general
principles also apply to a polyatomic molecule. Often the functional group in a
polyatomic molecule can be treated as a diatomic molecule (for example, )C=O
in a ketone or aldehyde). The electronic transition may occur in the functional
group between orbitals that are approximated by a combination of atomic orbitals
of the two atoms, as in a diatomic molecule. The actual energies of the resulting
molecular orbitals of the functional group will, of course, be affected by electronic,
conjugative, and steric effects arising from the other atoms. This situation can
be understood qualitatively in terms of potential energy curves similar to those
discussed for the diatomic molecules. For more complex cases in which several
atoms in the molecule are involved (i.e., a delocalized system), a polydimensional
surface is required to represent the potential energy curves.
The energies required for electronic transitions generally occur in the far
u.v., u.v., visible, and near infrared regions of the spectrum, depending upon the
energies of the molecular orbitals in a molecule. For molecules that contain only
strong sigma bonds (e.g., CH 4 and H 2 O), the energy required for electronic
transitions occurs in the far u.v. region of the spectrum and requires specialized
instrumentation for detection. In fact, in the selection of suitable transparent
solvents for study of the u.v. spectrum of the dissolved solute, such transitions
come into play in determining the cutoff point of the solvent. On the other hand,
the colored dyes used in clothing have extensively conjugated pi-systems and
undergo electronic transitions in the visible region of the spectrum. Standard
instruments cover the region from 50,000 cm-' to 5,000 cm-'. This spectral
region is subdivided as follows:
Ultraviolet 50,000 cm-1 to 26,300 cm-' (2000 to 3800 A)

Visible 26,300 cm-' to 12,800 cm-' (3800 to 7800 A)
Near infrared 12,800 cm-' to 5,000 cm- (7800 to 20,000 A)
5-3 Nomenclature 113
5-3 NOMENCLATURE
In our previous discussion we were concerned only with transitions of an electron

from a given ground state to a given excited state. In an actual molecule there
are electrons in different kinds of orbitals (a bonding, non-bonding, 7Tbonding)
with different energies in the ground state. Electrons from these different orbitals
can be excited to higher-energy molecular orbitals, giving rise to many possible
excited states. Thus, many transitions from the ground state to different excited
states (each of which can be described by a different potential energy curve) are
possible in one molecule.
There are several conventions(2 3, ) used to designate these different electronic
transitions. A simple representation introduced by Kasha(') will be illustrated for
the carbonyl group in formaldehyde. A molecular orbital description of the valence
electrons in this molecule is:
2 (U0 )
a2 72 n *)O
(n2
The n. and nb orbitals are the two non-bonding molecular orbitals containing
the lone pairs on oxygen. Symmetry considerations do not require the lone pairs E
to be degenerate, for there are no doubly degenerate irreducible representations N n,1)
in C 2 . They are not accidentally degenerate either, but differ in energy. The E
relative energies of these orbitals are indicated in Fig. 5-4. The ordering of these ' nj
orbitals can often be arrived at by intelligent guesses and by looking at the spectra Y
of analogous compounds. Also indicated with arrows are some transitions chosen
to illustrate the nomenclature. The transitions (1), (2), (3), and (4) are referred to
as n -> T*, n -- a*, m ->+n*, and o -* a*, respectively. The n -- 7r* transition is the Fie of elative
lowest energy-highest wavelength transition that occurs in formaldehyde and er ote carbo
most carbonyl compounds.
Electron excitations can occur with or without a change in the spin of the
electron. If the spin is not changed in a molecule containing no unpaired electrons,
both the excited state and ground state have a multiplicity of one; these states
are referred to as singlets. The multiplicity is given by two times the sum of the
individual spins, m, plus one: 2S + 1 = 21m, + 1. If the spin of the electron is
changed in the transition, the excited state contains two unpaired electrons with
identical magnetic spin quantum numbers, has a multiplicity of 3, and is referred
to as a triplet state.
There are some shortcomings of this simple nomenclature for electronic
transitions. It has been assumed that these transitions involve a simple transfer
of an electron from the ground state level to an empty excited state level of our
ground state wave function. For many applications, this description is precise
enough. In actual fact, such transitions occur between states, and the excited
state is not actually described by moving an electron into an empty molecular
orbital of the ground state. The excited state has, among other things, different
electron-electron repulsions than those in the system represented by simple
excitation of an electron into an empty ground state orbital. In most molecules,
various kinds of electron-electron interactions in the excited state complicate the
problem; in addition to affecting the energy, they give rise to many more
transitions than predicted by the simple picture of electron promotion because
levels that would otherwise be degenerate are split by electronic interactions.
This problem is particularly important in transition metal ion complexes. We
Chapter 5 Electronic Absorption Spectroscopy
shall return to discuss this problem more fully in the section on configuration
interaction.
In a more accurate(4 ) system of nomenclature the symmetry, configuration,
and multiplicity of the states involved in the transition are utilized in describing
the transition. This system of nomenclature can be briefly demonstrated by again
considering the molecular orbitals of formaldehyde. The diagrams in Fig. 5-5
qualitatively represent the boundary contours of the molecular orbitals. The solid
line encloses the positive lobe and the dashed line the negative lobe. The larger
7r and nr* lobes indicate lobes above the plane of the paper, and the smaller ones
represent those below the plane; the two lobes actually have identical sizes. To
classify these orbitals it is necessary first to determine the overall symmetry of
the molecule, which is C2,. The next step is to consult the C2 . character table.
FIGURE 5-5 Shapes of the N

molecular orbitals of 7 Ir*
formaldehyde. / I , I
nb n,
Character tables for several common point groups are listed in Appendix A. The
C2 . character table is duplicated here in Table 5-1. By convention, the yz-plane
is selected to contain the four atoms of formaldehyde. The symmetry operations
E, C 2 , and uv'(Yz) performed on the 71 orbital produce the result + 1, -1,
xv(XZ)
+ 1, - 1. This result is identical to that listed for the irreducible representation

B1 in the table. The orbital is said to belong to (or to transform as) the symmetry
species bi, the lowercase letter being employed for an orbital and the uppercase
TABLE 5-1. Character Table for the C2 v Point

Groupa
E C2 (Yr)
A1 1 1 1 1 z
A2 1 1 -1 1 R
B1 1 -1 1 -1 x,RY
B2 1 1 - yR
" The x-axis is perpendicular to the plane of the
molecule.
5,A Nomenclature 115
letters being reserved to describe the symmetry of the entire ground or excited
state. Similarly, if the n., n, it, o, and a* orbitals of formaldehyde are subjected
to the above symmetry operations, it can be shown that these orbitals belong to
the irreducible representations a1, b, b, al, and a, respectively. The two n
orbitals can be viewed as s and p, orbitals (p. is used in the a bond and p, in
the ). As a result, they lie in the yz-plane and possess a1 and b2 symmetry. (The
a1 s-orbital can mix with p. in the sigma bonding.) The difference in s-orbital
character causes the energies of the a1 and b2 lone pairs to differ. Molecular
orbital calculations are consistent with these ideas. As mentioned in Chapter 2,
a and b indicate single degeneracy. The a representation does not change sign
with rotation about the n-fold axis, but b does.
The symmetry species of a state is the product of the symmetry species of each
of the odd electron orbitals. In the state that results from the n -+ n* transition
in formaldehyde, there is one unpaired electron in the n-orbital with b 2 symmetry
and one in the 7r*-orbital with b, symmetry. The direct product is given by:
E C2 Uv(xz) av (yz)
b, x b2 = (1)(1) (-1)(- 1) (,)(- 1) (- 1)(1)
result = +1 +1 -1 -1 -A 2
The resulting irreducible representation is A 2 .* The excited state from this

transition is thus described as A 2 and the transition as A 1 -- A 2 . A common
convention involves writing the high energy state first and labeling the transition
A 2 +- A 1 . The spin multiplicity is usually included, so the complete designation
becomes 1A2 +- 'Al. The ground state is A1 because there is a pair of electrons
in each orbital. Commonly, the orbitals involved in the transitions are indicated,
and the symbol becomes 'A 2(n, 7r*) +- 'A 1 . If a non-specific, general symbol is
needed for a state symmetry species, F is employed.
Instead of representing the orbitals of formaldehyde symbolically, as in Fig.
5-5, we could simply have been given the wave functions. We can deduce the
symmetry from f by converting the wave functions into a physical picture. The
following equations describe the formaldehyde 7n and 7r* orbitals:
= ap, + bpp
b'<p, - a'<p
b.
where <p- and <p, are the wave functions for the atomic oxygen and carbon
p-orbitals, respectively. The atomic orbitals are mathematically combined to
produce the 7r and i* orbitals. Since oxygen is more electronegative (i.e., a > b),
it becomes clear why it is often stated that an electron is transferred from oxygen
to carbon in the n - +t* transition. We shall shortly return to a discussion of the
experimental spectrum of formaldehyde.
* The actual procedure for determining the symmetry of this state involves multiplication of the
orbital symmetries for all the electrons in the carbonyl group. However, all filled orbitals contain
two electrons whose product must be A. Thus, paired electrons have no effect on the final result
for the excited state symmetry.
Some of the molecular orbitals for benzene are represented mathematically

and pictorially in Fig. 5-6. Notice that a difference in sign between adjacent
atomic orbitals of the wave function represents a node (point of zero probability)
in the molecular orbital. Using the D6 1 character table, the symmetries of the
orbitals can be shown to be a2 ., ei,, e2., and b2, for 0r, 0 2 + 0 3 , 0 4 + 5 and
06, respectively.
In addition to the above conventions used to label ground and excited states,
a convention used for diatomic molecules will be described. With this terminology
the electronic arrangement is indicated by summing up the contributions of the
separate atoms to obtain the net orbital angular momentum. If all electrons are
paired, the sum is zero. Contributions are counted as follows: one unpaired
electron in a a orbital is zero, one in a 7r orbital one, and one in a 6 orbital two.
For more than one electron, the total is IIm. If the total is zero, the state is
described as I, one as fl, and two as A. The multiplicity is indicated by a
superscript, e.g., the ground state of NO is 2 FI. A plus or minus sign often follows
the symbol to illustrate, respectively, whether the molecular orbitals are symmetric
or antisymmetric to a plane through the molecular axis.
a
fC - cCb
I I
eC Cc
Lettering sequence for a.o.'s d
a2 1 / + r, + T + Td + p, + ,) zero nodes
Degenerate,
e,< >set -
one node
2
0P3 (1/ T,
) (q),+C 9-,-),
14 = (1 /2) (1b W+Degenerate
e2.< > set-

2
two nodes
'P5 1hITl(2T% We+ 9'd -e Wf)
b29.,0 1A6)(a-9 +9 -Td+T 9 Three nodes
FIGURE 5-6 Shapes of the benzene molecular orbitals.

5-5 Configuration Interaction 117
If the energies of electronic transitions were related to the ground state Assignment of
molecular orbital energies, the assignment of transitions to the observed bands
would be simple. In formaldehyde (Fig. 5-4) the n --+7* would be lower in energy Transitions
than the 7r -+ 7t*, which in turn would be lower in energy than - - n*. In addition
to different electron-electron repulsions in different states, two other effects
complicate the picture by affecting the energies and the degeneracy of the various
excited states. These effects are spin-orbit coupling and higher-state mixing.
5-4 SPIN-ORBIT COUPLING
There is a magnetic interaction between the electron spin magnetic moment

(signified by quantum number m, = %'/2)and the magnetic moment due to the
orbital motion of an electron. To understand the nature of this effect, consider
the nucleus as though it were moving about the electron (this is equivalent to
being on earth and thinking of the sun moving across the sky). We consider the
motion from this reference point because we are interested in effects at the electron.
The charged nucleus circles the electron, and this is equivalent in effect to placing
the electron in the middle of a coil of wire carrying current. As moving charge
in a solenoid creates a magnetic field in the center, the orbital motion described
above causes a magnetic field at the electron position. This magnetic field can
interact with the spin magnetic moment of the electron, giving rise to spin-orbit
interaction. The orbital moment may either complement or oppose the spin
moment, giving rise to two different energy states. The doubly degenerate energy
state of the electron (previously designated by the spin quantum numbers ± '/)
is split, lowering the energy of one and raising the energy of the other. Whenever
an electron can occupy a set of degenerate orbitals that permit circulation about
the nucleus, this interaction is possible. For example, if an electron can occupy
the d,, and dxz orbitals of a metal ion, it can circle the nucleus around the z-axis.
(A more complete discussion will be given in Chapter 11.)
5-5 CONFIGURATION INTERACTION
As mentioned before, electronic transitions do not occur between empty molecular

orbitals of the ground state configuration, but between states. The energies of
these states are different from those of configurations derived by placing electrons
in the empty orbitals of the ground state, because electron-electron repulsions in
the excited state differ from those in the simplified, "ground state orbital
description" of the excited state. A further complication arises from configuration
interaction. One could attempt to account for the different electron-electron
repulsions in the excited state by doing a molecular orbital calculation on a
molecule having the ground state geometry but with the electron arrangement
of the excited configuration. This would not give the correct energy of the state
because this configuration can mix with all of the other configurations in the
molecule of the same symmetry by configuration interaction.
This mixing is similar in a mathematical sense (though much smaller in
magnitude) to the interaction of two hydrogen atoms in forming the H2 molecule.
= (1 - a 2) 1 12 ,B - ( b 2 )1 2
4B
'
-B j
Before --- BBi + bPB

=apBa
C.- After C.I.
FIGURE 5-7 Energy levels before and after configuration

interaction.
Thus, we could write a secular determinant to account for the mixing of two B 1
states, B'ia and B'ib, as:
Ei -E H1 2 _ 0
H12 E20 _ E
where H 1 2 is the difficult-to-solve integral jf(Bia)fII(Bib) dr, whose value is
dependent upon the interelectron repulsions in the various states. The closer in
energy the initial states E,' and E 20 , the more mixing occurs. Solution of the
secular determinant gives us the two new energies after mixing:
Ei = [E 1 + E 2 0 + ((E1 0)2 + (E2 0)2 2E 10 E 2 0 + 4H 1 2 2 )1/2]

E2 =[E10 + E2 0 - ((E1 0 )2 + (E2 0)2 - 2E 0 0 2 1/
1 E2 + 4H1 2 ) 2]
The energies of the initial (B'ia, B'i,) and final (Bia, Bib) states are illustrated in
Fig. 5-7 along with the new wave functions to describe the final state. The wave
function is seen to be a linear combination of the two initial molecular orbitals.
Interactions of this sort can occur with all the molecular orbitals of B, symmetry
in the molecule, complicating the problem even further than is illustrated in Fig.
5-7.
5-6 CRITERIA TO AID IN BAND ASSIGNMENT
An appreciation for some of the difficulties encountered in assigning transitions

5
can be obtained from reading the literature( ) and noting the changes in the
assignments that have been made over the years. Accordingly, many independent
criteria are used in making the assignment. These include the intensity of the
transition and the behavior of the absorption band when polarized radiation is
employed. Both of these topics will be considered in detail shortly. Next we shall
describe some simple observations that aid in the assignment of the n -- 7E* and
7 * transitions.
->
For n -i n* transitions, one observes the following characteristics:
1. The molar absorptivity of the transition is generally less than 2000. An
explanation for this is offered in the section on intensity.
2. A blue shift (hypsochromic shift, or shift toward shorter wavelengths) is
5-6 Criteria to Aid in Bond Assignment 119
observed for this transition in high dielectric or hydrogen-bonding solvents. This

indicates that the energy difference between the ground and excited state is
increased in a high dielectric or hydrogen-bonding solvent. In general, for solvent
shifts it is often difficult to ascertain whether the excited state is raised in energy
or the ground state lowered. A blue shift may result from a greater lowering of
the ground state relative to the excited state or a greater elevation of the excited
state relative to the ground state. It is thought that the solvent shift in the n - n*
transition results from both a lowering of the energy of the ground state and an
elevation of the energy of the excited state. In a high dielectric solvent the molecules
arrange themselves about the absorbing solute so that the dipoles are properly
oriented for maximum interaction (i.e., solvation that lowers the energy of the
ground state). When the excited state is produced, its dipole will have an
orientation different from that of the ground state. Since solvent molecules cannot
rearrange to solvate the excited state during the time of a transition, the excited
state energy is raised in a high dielectric solvent.(6)
Hydrogen bonding solvents cause pronounced blue shifts. This is reported
to be due to hydrogen bonding of the solvent hydrogen with the lone pair of
electrons in the n orbital undergoing the transition. In the excited state there is
only one electron in the n orbital, the hydrogen bond is weaker and, as a result,
the solvent does not lower the energy of this state nearly as much as that of the
ground state. In these hydrogen bonding systems an adduct is formed, and this
specific solute-solvent interaction is the main cause of the blue shift.(') If there is
more than one lone pair of electrons on the donor, the shift can be accounted
for by the inductive effect of hydrogen bonding to one electron pair on the energy
of the other pair.
3. The n - 7r* band often disappears in acidic media owing to protonation
or upon formation of an adduct that ties up the lone pair, e.g., BCH+ I (as in
CHNCHI ), where B is the base molecule containing the n electrons. This
behavior is very characteristic if there is only one pair of n electrons on B.
4. Blue shifts occur upon the attachment of an electron-donating group
to the chromophore. For example in the series
CH 3C(O)H < CH 3C(O)CH 3 < CH 3C(O)OCH 3 < CH 3C(O)N(CH 3)2
an increasing blue shift is observed in the carbonyl absorption band. A molecular

orbital treatment(') indicates that this shift results from raising the excited 7E*
level relative to the n level.
5. The absorption band corresponding to the n -+*r* transition is absent
in the hydrocarbon analogue of the compound. This would involve, for example,
comparison of the spectra of benzene and pyridine or of H2 C=O and H2 C=CH2 .
6. Usually, but not always, the n -* n* transition gives rise to the lowest
energy singlet-singlet transition.
In contrast to the above behavior, 7r - ir* transitions have a high intensity.
A slight red (bathochromic) shift is observed in high dielectric solvents and upon
introduction of an electron-donating group. It should be emphasized that in the
above systems only the difference in energy between the ground and excited states
can be measured from the frequency of the transition, so only the relative energies
of the two levels can be measured. Other considerations must be invoked to
determine the actual change in energy of an individual state.
The Intensity 5-7 OSCILLATOR STRENGTHS

of Electronic As mentioned in the previous chapter, the intensity of an absorption band can be
Transitions indicated by the molar absorptivity, commonly called the extinction coefficient. A
parameter of greater theoretical significance is f, the oscillator strength of
integrated intensity, often simply called the integrated intensity:
f= 4.315 x 10-' f edv (5-1)
In equation (5-1), E is the molar absorptivity and 9 is the frequency expressed in

wave numbers. The concept of oscillator strength is based on a simple classical
model for an electronic transition. The derivation* indicates that f = 1 for a fully
allowed transition. The quantity f is evaluated graphically from equation (5-1)
by plotting e, on a linear scale, versus the wavenumber P in cm -1, and calculating
the area of the band. Values of f from 0.1 to 1 correspond to molar absorptivities
in the range from 10,000 to 100,000, depending on the width of the peak.
For a single, symmetrical peak, f can be approximated by the expression:
9
f ; (4.6 x 10 )e. Av, 2 (5-2)
where ea is the molar absorptivity of the peak maximum and Av,/2 is the half
intensity bandwidth, i.e., the width at E./2.
5-8 TRANSITION MOMENT INTEGRAL
The integrated intensity, f, of an absorption band is related to the transition

moment integral as follows:
fcoc J + o2
f eie,,,x dv =D (5-3)
where D is called the dipole strength, fe and feix are electronic wave functions
for the ground and excited states respectively, Mi is the electric dipole moment
operator (vide infra), and the entire integral is referred to as the transition moment
integral. To describe *, one should recall that the electric dipole moment is
defined as the distance between the centers of gravity of the positive and negative
charges times the magnitude of these charges. The center of gravity of the positive
charges in a molecule is fixed by the nuclei, but the center of gravity of the
electrons is an average over the probability function. The vector for the average
distance from the nuclei to the electron is represented as r. The electric dipole
moment vector, M, is given by M = Ee, with the summation carried out over
* For this derivation, see Additional References, Barrow, pages 80 and 81.
5-8 Transition Moment Integral 121
all the electrons in the molecule. For the ground state, the electric dipole moment
is given by the average over the probability function, or
f V/gYe i, dT
By comparison of this equation for the ground state dipole moment with the
transition moment integral:
S01 rj 1 fex dz
this integral can be seen roughly to represent charge migration or displacement

during the transition.
When the integral in equation (5-3) is zero, the intensity will be zero and,
to a first approximation, the transition will be forbidden. In general, we do not
have good state wave functions to substitute into this equation to calculate the
intensity, for reasons discussed in the previous section. However, symmetry can
tell us if integrals are zero. It is important to examine the symmetry properties
of the integrand of equation (5-3), for this enables us to make some important
predictions. These symmetry considerations will also lead to selection rules for
electronic transitions. The quantity A is a vector quantity and can be resolved
into x, y, and z components. The integral in equation (5-3) then has the
components:
fe k/ei** dv (5-4)
I dv
0,c1*Y0ex (5-5)
T 0IAel4zoe*x dv (5-6)
In order to have an allowed transition, at least one of the integrals in equations

(5-4) to (5-6) must be non-zero. If all three of these integrals are zero, the transition
is called forbidden and, according to approximate theory, should not occur at
all. Forbidden transitions do occur; more refined theories (discussed in the section
on Spin-Orbit and Vibronic Coupling) give small values to the intensity integrals.
Symmetry considerations can tell us if the integral is zero in the following
way. An integral can be non-zero only if the direct product of the integrand
belongs to symmetry species A 1. Another way of saying this is that such integrals
are different from zero only when the integrand remains unchanged for any of
the symmetry operations permitted by the symmetry of the molecule. The reason
for the above statements can be seen by looking at some simple mathematical
functions. First, consider a plot of the curve y = x, shown in Fig. 5-8(A). Symmetry
operations of the C,, point group on this figure tell us that this function does
not have A 1 symmetry. The integral fy(x) dx represents the area under the curve.
Since y(x) is positive in one quadrant and negative in the other, the area from
+ a to - a or from + oo to - oo is zero. Now, plot a function y(x) such that y

is symmetric in x. One of many such functions is y = x2 , shown in Fig. 5-8(B).
This function is called "even" and does have A 1 symmetry, because it is
unchanged by any of the operations of the symmetry group to which the function
belongs. This can be shown by carrying out the symmetry operations of the C 2,
point group on Fig. 5-8(B). The integral does not vanish. The shaded area gives
the value for the integral ja a y dx for y = x 2, and it is seen not to cancel to zero.
Since y is a function of x 2 , we can take the direct product of the irreducible
representation of x with itself, Fx x Fx, and reduce this to demonstrate that this
is a totally symmetric irreducible representation. Next, the function y = x3, in
Fig. 5-8(C), will be examined. Symmetry considerations tell us that this is an odd
function, and we see that jy dx (for y = x 3) = 0. Does Fx x Fx x Fx = A 1 ? The
integral over all space of an odd function always vanishes, as shown for the
examples in Figs. 5-8(A) and 5-8(C). The integral over all space of an even
function generally does not vanish.
y =x
-a a
(A) (B) (C)

FIGURE 5-8 Plots of some simple functions, y = f(x), and their symmetries.
Thus, to determine whether or not the integral is zero, we take the direct
product of the irreducible representations of everything in the integrand. If the
direct product is or contains A1, the integral is non-zero and the transition is
allowed.
The application of these ideas is best illustrated by treating an example.
Consider the 7r -+ nr* transition in formaldehyde. The ground state, like all ground
states containing no unpaired electrons, is A. The excited state is also
A 1 (b, x b, = A 1 ). The components k., A,, and k. transform as the x, y, and
z-vectors of the point group. The table for the C2,, point group indicates that
M, the dipole moment vector lying along the z-axis, is A. Since
A1 x A, x A 1 = A, the integrand 0 A,4/ k for the 7 - * is Ai, and the n - *
transition is allowed.
For an n -+ 7r* transition, the ground state is A, and the excited state is A 2 -
The character table indicates that no dipole moment component has symmetry
A 2 . Therefore, none of the three integrals [equations (5-4) to (5-6)] can be A1,
and the transition is forbidden (A2 x A 2 is the only product of A 2 that is A1 ).
As was mentioned when we introduced this topic, the transition moment
integral can be used to derive some important selection rules for electronic
transitions.
5-10 Spectrum of Formaldehyde 123
5-9 DERIVATION OF SOME SELECTION RULES
1. For molecules with a center of symmetry, allowed transitions are g --+ u

or u - g. (The abbreviations g and u refer to gerade and ungerade, which are
German for even and odd, respectively.) The d- and s-orbitals are g, and p-orbitals
are u. All wave functions in a molecule with a center of symmetry are g or u. All
components of the vector ft in a point group containing an inversion center are
necessarily ungerade.
Fog x , x F0, = F
u x u x u= u forbidden
u x u x g = g allowed
g x u x g =u forbidden
g x u > u =g allowed
This leads to the selection rule that g -+ u and u -+ g are allowed, but g - g and
u -+ u are forbidden. Therefore, d -- d transitions in transition metal complexes
with a center of symmetry are forbidden. Values of e for the d-d transitions in
Ni(H 2 O), 2 + are ~20.
2. Transitions between states of different multiplicity are forbidden. Con-
sider a singlet -- triplet transition. Focusing on the electron being excited, we
have in the singlet ground state 0faffl and, in the excited state, coafrc
or f# #,
where a and # are the spin coordinates. The dipole strength is given by
2
D= f iMjcd f# dT do-
(where da is the volume element in the spin coordinates and the i and f subscripts
refer to initial and final states). We can rewrite the integral corresponding to D as
1 dz f ca# da
jMoi
Since the second term is the product of + 1/2 and - 1/2 spins, it is always odd
and zero, i.e., the spins are orthogonal. The E for absorption bands involving
transitions between states of different multiplicity is generally less than one. Since
fjca do = 1 and ff# do = 1, in working out the intensity integral we only have
to worry about the electron that is undergoing the transition, and we can ignore
all the electrons in the molecule that do not change spin.
3. Transitions in molecules without a center of symmetry depend upon the
symmetries of the initial and final states. If the direct product of these and any
one of k,, k,, or 142 is A,, the transition is allowed. If all integrals are odd,
the transition is forbidden.
5-10 SPECTRUM OF FORMALDEHYDE
We can summarize the above ideas and illustrate their utility by returning
again to the ultraviolet spectrum of formaldehyde. The various possible excited
states arising from electron excitations from the highest-energy filled orbitals
(na, no, and 7r) are given by:
a,2 b12 b21 bl* = 'A 2 (nb 7r*)
al 2 b 2b2la * = 'B 2 (nb -6*)

2
a, bi1b 2
2
b l* = 'A, (7[ *i*)
2 1
a, b, b2 a,'* = 1B,
2
(nZ *6*)
alibi2 b22b* = 'B, (na 7 *)
ab 2b 1a = 'A, (na U*)
Two bands are observed, one with e = 100 at 2700 A and an extremely intense
one at 1850 A. We see from Fig. 5-4 that the lowest-energy transitions are nb - *
and 71 -- * nr*, better expressed as 'A, -* 'A 2 and 'A, -+ 'A,. The
'A1 --+ A 2(nb IT*) is forbidden, and accordingly is assigned to the band at
-+
2700 A. Both 'A, - 'B1(na -+ nr*) and 'A, -+ 'A,(n - i*) are allowed. The former
may contribute to the observed band at 1850 A or may be in the far u.v.
The integrands for 'A, -+ 'B,(nt-*o*), 'A, -* 'B 2(nb -+ a*), and
'A, -* 'AI(na - a*) are all A, leading to allowed transitions. These are expected
to occur at very short wavelengths in the far ultraviolet region. This is the presently
held view of the assignment of this spectrum, and it can be seen that the arguments
are not rigorous. We shall subsequently show how polarization studies aid in
making assignments more rigorous.
Next, it is informative to discuss the u.v. spectrum of acetaldehyde, which
is quite similar to that of formaldehyde. The nb ' 7* transition has very low
intensity. However, acetaldehyde has C, symmetry; this point group has only
two irreducible representations, A and B, with the x- and y-vectors transforming
as A and the z-vector as B. Accordingly, all transitions will have an integrand
with A, symmetry and will be allowed. Though the nb - * transition is allowed
by symmetry, the value of the transition moment integral is very small and the
intensity is low. The intensity of this band in acetaldehyde is greater than that
in formaldehyde. We can well appreciate the fact that although monodeuterofor-
maldehyde [DC(O)H] does not have C,, symmetry, it will have an electronic
spectrum practically identical to that of formaldehyde. These are examples of a
rather general type of result, which leads to the idea of local symmetry. According
to this concept, even though a molecule does not have the symmetry of a particular
point group, if the groups attached to the chromophore have similar bonding
interactions, (e.g. CH 3 and C 2H,) the molecule for many purposes can be treated
as though it had this higher symmetry.
5-11 SPIN-ORBIT AND VIBRONIC COUPLING

CONTRIBUTIONS TO INTENSITY
The discrepancy between the theoretical prediction that a transition is forbidden

and the experimental detection of a weak band assignable to this transition is
attributable to the approximations of the theory. More refined calculations that
include effects from spin-orbit coupling often predict low intensities for otherwise
5-11 Spin-Orbit and Vibronic Coupling Contributions to Intensity 125
forbidden transitions. For example, a transition between a pure singlet state and
a pure triplet state is forbidden. However, if spin-orbit coupling is present, the
singlet could have the same total angular momentum as the triplet and the two
states could interact. The interaction is indicated by equation (5-7):
= af + b 30 (5-7)
where i and '0 correspond to pure singlet and triplet states, respectively, /
represents the actual ground state, and a and b are coefficients indicating the
relative contributions of the pure states. If a > b, the ground state is essentially
singlet with a slight amount of triplet character and the excited state will be
essentially triplet. This slight amount of singlet character in the predominantly
triplet excited state leads to an intensity integral for the singlet-triplet transition
that is not zero; this explains why a weak peak corresponding to the multiplicity-
forbidden transition can occur.
Another phenomenon that gives intensity to some forbidden transitions is
vibronic coupling. We have assumed until now that the wave function for a
molecule can be factored into an electronic part and a vibrational part, and we
have ignored the vibrational part. When we applied symmetry considerations to
our molecule, we assumed some symmetrical, equilibrium internuclear configu-
ration. This is not correct, for the molecules in our system are undergoing
vibrations and during certain vibrations the molecular symmetry changes. For
example, in an octahedral complex, the T, and Tu vibrations shown in
Fig. 5-9 remove the center of symmetry of the molecule. Since electronic
transitions occur much more rapidly than molecular vibrations, we detect
transitions occurring in our sample from many geometries that do not have high
symmetry, e.g., the vibrationally distorted molecules of the octahedral complex
shown in Fig. 5-9. The local symmetry is still very close to octahedral, so the
intensity gained this way is not very great; but it is large enough to allow a
forbidden transition to occur with weak intensity.
The electronic transition can become allowed by certain vibrational modes
but not by all. We can understand this by rewriting the transition moment integral
to include both the electronic and the vibrational components of the wave function
as in equation (5-8):
f x D = f OeI4IibMoIetvib*e dbr (5-8)
As we mentioned in Chapter 4, all ground vibrational wave functions are A1, so

the symmetry of 0feIvib becomes that of 0e, which is also A 1 for molecules with
no unpaired electrons. (In general discussion, we shall use the symbol A, to
represent the totally symmetric irreducible representation, even though this is
not the appropriate label in some point groups.) To use this equation to see
whether a forbidden transition can gain intensity by vibronic coupling, we must
take a product i that is not A, and see whether there is a vibrational
*(xyor*z)oe1
mode with symmetry that makes the product MIX equal to A.
*Y'*)0o"0vibex
When t,/'** has the same symmetry as the product M( the product
**),ex,
will be A I.
This discussion can be made clearer by considering some examples. We shall
consider vibrational spectroscopy in more detail in the next chapter. A non-linear
FIGURE 5-9 T1. and T2.

vibrations of an octahedral
complex. \
molecule has 3N - 6 internal vibrations; for formaldehyde these are 3ai, b, and
2b 2 . For the forbidden transition 'A, -- 'A 2(n, - n*), the vibrational wave
function of al symmetry does not change the direct product ]I(XOrZ) 'A 2 so no
intensity can be gained by this mode. Excitation of the b, vibrational mode leads
to a direct product 0'f*re ib*e of b, x A 2 = B 2. Since M(, has B 2 symmetry, the
total integral (flPe,1vibpyfeexofi**bex dr) has A, symmetry, and the electronic
transition becomes allowed by vibronic coupling to the b, mode.
It is informative to consider Co(NH 3)6 3 + as an example, for it contains
triply degenerate irreducible representations. The ground state is 'A g (a strong
field 0 h d' complex). The excited states from d-d transitions are 'T and
1 T g- M, A,, and A, transform as T,.. For the 'A g - 'T, transition one
2
obtains:
Alg X Ti. X Ti
The resulting direct product representation has a dimensionality of nine (the

identity is 1 x 3 x 3 = 9) and the total representation is reduced into a linear
combination of A,. + E. + T1. + T2 . irreducible representations. With no A g
component, the 'A , - T,, transition is forbidden. However, the vibrations for
an octahedral complex have the symmetries of ag, eg, 2t,., t 2 g, t2u. Since the
direct products t,, x T,. and t2. x T2. have Alg-components, this transition
becomes allowed by vibronic coupling. For practice, the reader should take the
direct products and factor the reducible representations discussed here.
5-12 MIXING OF d AND p ORBITALS IN CERTAIN

SYMMETRIES
There is one further aspect of the intensity of electronic transitions that can be
understood via the symmetry aspects of electronic transitions. The electronic
spectra of tetrahedral complexes of cobalt(II) contain two bands assigned to d-d
transitions at ~ 20,000 cm- 'and ~6000 cm-', assigned as A 2 -+ T, and A 2 -+ T2
transitions respectively, with molar absorptivities of 600 and 50. Since the
-components transform as T2 , we obtain for the A 2 --+ T, transition
A 2 x T2 x T1 = A, + E + T + T2
5-14 Change Transfer Transitions 127
so the transition is allowed. However, if only the d-orbitals were involved in this
transition, the intensity would be zero for the integrals
f O .,y Mlfrd xz d r= 0
However, in the T point group, the d,, d., and d,, orbitals and the p orbitals
transform as T2 and therefore can mix. If the two states involved in the transition,
A 2 and T, have differing amounts of p-character, intensity is gained by having
some of the highly allowed p -- d or d -* p character associated with the transition.
Consider the consequences of this mixing on the A2 --+ T transition. The
transition moment integrand for this transition is
A 2 x T2 x T2
which, as the reader should verify, can be reduced to A 2 + E + Ti + T2 . Since

there is no A, component, the transition is forbidden. Mixing p-character into
the wave functions will not help, for this type of transition is still forbidden.
Accordingly, the r for the A 2 -- T, transition is ten times greater than that of
A 2 -4 T2 . The latter transition gains most of its intensity by vibronic coupling.
5-13 MAGNETIC DIPOLE AND ELECTRIC QUADRUPOLE

CONTRIBUTIONS TO INTENSITY
So far, our discussion of the intensity of electronic transitions has centered on

the electric dipole component of the radiation, with the transition moment integral
involving the electric dipole operator, e r. There is also a magnetic dipole
component. The magnetic dipole operator transforms as a rotation R_,RR,, and
the intensity from this effect may be regarded as arising from the rotation of
electron density. Transition moment integrals similar to those for electric dipole
transitions can be written for the contribution from both magnetic dipole and
electric quadrupole effects. In a molecule with a center of symmetry, both of these
operators are symmetric with respect to inversion, so g - g and u -u transitions
are allowed. Approximate values of the transition moment integral for allowed
transitions for these different operators are: 6 x 10-36 cgs units for an electric
dipole transition, 9 x 10-41 cgs units for a magnetic dipole transition, and
7 x 10 - cgs units for a quadrupole transition. Thus, we can see that these latter
two effects will be important only when electric dipole transitions are forbidden.
They do complicate the assignment of very weak bands in the spectrum.
5-14 CHARGE TRANSFER TRANSITIONS
A transition in which an electron is transferred from one atom or group in the

molecule to another is called a charge-transfertransition. More accurately stated,
the transition occurs between molecular orbitals that are essentially centered on
different atoms. Very intense bands result, with molar absorptivities of 104 or
greater. The frequency at maximum absorbancy, vm.x, often, but not always,
occurs in the ultraviolet region. The anions C10 4 - and SO,4 show very intense
2
bands. Since MnO 4 - and Cr0 4 have no d electrons, the intense colors of these
ions cannot be explained on the basis of d-d transitions; they are attributed to
9 2
charge-transfer transitions.! ) The transitions in MnO 4 - and Cr0 4 - are most
simply visualized as an electron transfer from a non-bonding orbital of an oxygen
atom to the manganese or chromium (n -+ r*), in effect reducing these metals in
the excited state." ) An alternate description for this transition involves excitation
of an electron from a 7r bonding molecular orbital, consisting essentially of oxygen
atomic orbitals, to a molecular orbital that is essentially the metal atomic orbital.
In the case of a pyridine complex of iridium(III), a charge-transfer transition
that involves oxidation of the metal has been reported."') A metal electron is
transferred from an orbital that is essentially an iridium atomic orbital to an
empty nt* antibonding orbital in pyridine.
In gaseous sodium chloride, a charge-transfer absorption occurs from the
ion pair Na*Cl- to an excited state described as sodium and chlorine atoms
having the same internuclear distance as the ion pair. A charge-transfer absorption
also occurs in the ion pair, N-methylpyridinium iodide(3 6 ) (see Fig. 5-13) in which
an electron is transferred from I- to a ring antibonding orbital. The excited state
is represented in Fig. 5-13. A very intense charge-transfer absorption is observed
in addition compounds formed between iodine and several Lewis bases. This
phenomenon will be discussed in more detail in a later section.
5-15 POLARIZED ABSORPTION SPECTRA
If the incident radiation employed in an absorption experiment is polarized, only

those transitions with similarly oriented dipole moment vectors will occur. In a
powder, the molecules or complex ions are randomly oriented. All allowed
transitions will be observed, for there will be a statistical distribution of crystals
with dipole moment vectors aligned with the polarized radiation. However,
suppose, for example, that a formaldehyde crystal, with all molecules arranged
so that their z-axes are parallel, is examined. As indicated in the previous section,
the integrand f*AIz has appropriate symmetry for the 'A(,.) +- 'A transition,
but 0*k.A and qf*yo do not. When the z-axes of the molecules in the crystal
are aligned parallel to light that has its electric vector polarized in the z-direction,
light will be absorbed for the 'A(,*) +- 'A transition. Light of this wavelength
polarized in other planes will not be absorbed. If this crystal is rotated so that
the z-axis is perpendicular to the plane of polarization of the light, no light is
absorbed. This behavior supports the assignment of this band to the transition
'A, +- 'A,. To determine the expected polarization of any band, the symmetry
species of the product /aIb is compared with the components of Api, as was done
before for formaldehyde. The polarization experiment is schematically illustrated
in Fig. 5-10.
In Fig. 5-10(A), absorption of radiation will occur if Mz results in an A,
transition moment integrand for equation (5-6). No absorption will occur if it is
not A, regardless of the symmetries of the integrand for the k or MY components
[i.e., equations (5-4) and (5-5)]. In Fig. 5-10B, absorption will occur if the M,
component gives an A, transition moment integral. Even if Mz has an integrand
with A, symmetry, no absorption of the z-component will occur for this
orientation and absorption will not occur unless the M, integrand is A,.
5-15 Polarized Absorption Spectra 129
Plane polarized oscillation z-axis of crystal

FIGURE 5-10 Schematic
Source
IL x
illustration of a polarized
single crystal study. (A) The
z-axis of the crystal is
parallel to the oscillating
Nicol prism electromagnetic plane
polarized component. (B) The
Plan e y z-axis of the crystal is
polariz ed perpendicular to the
oscillat ion
oscillating electromagnetic
z plane polarized component,
and the y-axis is parallel to it.
We can further illustrate these ideas by considering the electronic absorption

spectrum of PtC 4 -. The transitions are charge transfers involving electron
excitation from a mainly chlorine m.o. to the empty d,2 _,2 orbital on Pt(II). The
symmetry is D4,; using the same approach as that employed in Chapter 3 on the
NO 2 ion, on a basis set of four p. orbitals on chlorine, we obtain symmetry
orbitals for chlorine of b2 ., e., and a 2 . symmetry. This leads to the following
possible charge-transfer transitions:
b2.00) - b,(d,2 _,2) with state labels 'A g -+ lA 2 u (here A 2 . is the

direct product of b 2 . x b1g)
e.(n) - bg(d,2 _,2) with state labels 'Ag -+E.
a 2 .(7r) - blg(d,2 _,2) with state labels Aig -'B2.
In the D 4h point group, k. and k, transform as E., and M as A 2 u. If we

first consider the Ajg -* A 2 u transition, we get for k.:
AigA 2 uA2u = Ag
and for k. and k, we get:
AigEuA 2u A,,
Accordingly, this transition is allowed and is polarized in the z-direction. If we

use polarized light and a single crystal, light will be absorbed when the z-axis of
the crystal is parallel to the z-direction of the light; but there will be no absorption
when the z-axis is perpendicular to the light because the M. and M, integrands
are not Al.
For the band assigned to 'A~g -+ 'Eu, the AigEuEu product has an Al
component, so this transition is also allowed. Since Mz yields AlgA2uEu, which
does not have an Alg component, there will be no absorption when the
z-component is parallel to the plane of the polarized light but absorption will
occur when the x and y-axes of the crystal are parallel to the light.
The A g - B2 . transition turns out to be forbidden. Thus, we see that by

employing polarized single crystal spectroscopy, we can rigorously assign the
2
two intense charge-transfer bands observed in the electronic spectrum of PtCl -.
If the single crystal employed in these experiments did not have all of the molecular
z-axes aligned, the polarization experiments would not work.
Applications Most applications of electronic spectroscopy have been made in the wavelength
range from 2100 to 7500 A, for this is the range accessible with most recording
spectrophotometers. Relatively inexpensive commercial instruments can now be
obtained to cover the range from 1900 to 8000 A. The near infrared region, from
8000 to 25,000 A, has also provided much useful information. Spectra can be
examined through the 1900 to 25,000 A region on samples of vapors, pure liquids,
or solutions. Solids can be examined as single crystals or as discs formed by
mixing the material with KC1 or NaCl and pressing with a hydraulic press until
a clear disc is formed.(' 2 ) Spectra of powdered solids can also be examined over
a more limited region (4000 to 25,000 A) as reflectance spectra or on mulls of
12
the solid compounds.( )
5-16 FINGERPRINTING
Since many different substances have very similar ultraviolet and visible spectra,
this is a poor region for product identification by the "fingerprinting" technique.
Information obtained from this region should be used in conjunction with other
evidence to confirm the identity of the compound. Evidence for the presence of
functional groups can be obtained by comparison of the spectra with reported
data. For this purpose, V=., rmax, and band shapes can be employed. It is also
important that the spectra be examined in a variety of solvents to be sure that
the band shifts are in accord with expectations (see discussion of blue shifts).
Spectral data have been compiled by Sadtler (see Additional References),
5 6
Lang,(1 3 ) and Hershenson,(14 ) and in "Organic Electronic Spectral Data."" , ) A
1
review article by Mason(1 7) and the text by Jaffe and Orchin ) are excellent for
this type of application. If a functional group (chromophore) is involved in
conjugation or steric interactions, or is attached to electron-releasing groups, its
spectral properties are often different from those of an isolated functional group.
These differences can often be predicted semiquantitatively for molecules in which
such effects are expected to exist.(7)
The spectra of some representative compounds and examples of the effect
of substituents on the wavelength of a transition will be described briefly.
SATURATED MOLECULES
Saturated molecules without lone pair electrons undergo high-energy

a -+ a* transitions in the far ultraviolet. For example, methane has a maximum
at 1219 A and ethane at 1350 A corresponding to this transition. When lone pair
TABLE 5-2. Frequencies of Electronic Transitions in Some
Saturated Molecules
Compound ma A E. Medium
H2 0 1667 1480 vapor

MeOH 1835 150 vapor
Me 2O 1838 2520 vapor
Me 2 S 2290,2100 140,1020 ethanol
S8 2750 8000 ethanol
F2 2845 6 vapor
Cl2 3300 66 vapor
Br2 4200 200 vapor
12 5200 950 vapor
I0 -4600 153 CC14
SC2 3040 1150 CC14
P13 3600 8800 Et 2O
Asl 3 3780 1600 pet. ether
electrons are available, a lower-energy n -> a* transition is often detected in

addition to the a --+ o*. For example, in triethylamine two transitions are observed,
at 2273 and 1990 A.
Table 5-2 contains a listing of absorption maxima for some saturated
compounds and gives some indication of the variation in the range and intensity
of transitions in saturated molecules.
CARBONYL COMPOUNDS
The carbonyl chromophore has been very extensively studied. Upon

conjugation of the carbonyl group with a vinyl group, four n1energy levels are
formed. The highest occupied 7r level has a higher energy, and one of the lowest
empty 7r* levels has a lower energy, than the corresponding levels in a
nonconjugated carbonyl group. The lone pair and a electrons are relatively
unaffected by conjugation. As a result, the n --+n* and n --+ 7t* transition energies
are lowered and the absorption maxima are shifted to longer wavelengths when
the carbonyl is conjugated. The difference is greater for the 7r -+ 7* than for the
n --+t* transition. The n -+ o* band is not affected appreciably and often lies
beneath the shifted 71 ->+t* absorption band. As stated earlier, electron-donating
groups attached to the carbonyl cause a blue shift in the n - n* transition and
a red shift in n -+ i*.
It is of interest to compare the spectra of thiocarbonyl compounds with
those of carbonyl compounds. In the sulfur compounds, the carbon-sulfur 7t
interaction is weaker and, as a result, the energy difference between the 7t and
it*-orbitals is smaller than in the oxygen compounds. In addition, the ionization
potential of the sulfur electrons in the thiocarbonyl group is less than the ionization
potential of oxygen electrons in a carbonyl. The n electrons are of higher energy
in the thiocarbonyl and the n -- n* transition requires less energy in these
compounds than in carbonyls. The absorption maximum in thiocarbonyls occurs
at longer wavelengths and in some compounds is shifted into the visible region.
INORGANIC SYSTEMS
The SO2 molecule has two absorption bands in the near ultraviolet at
3600 A (r = 0.05) and 2900 A (E= 340) corresponding to a triplet and singlet
n - n* transition. The gaseous spectrum shows considerable vibrational fine
structure, and analysis has produced information concerning the structure of the
excited state.(1 8)
In nitroso compounds, an n -- * transition involving the lone pair electrons
on the nitrogen occurs in the visible region. An n -- 7E*transition involving an
oxygen lone pair occurs in the ultraviolet.
The nitrite ion in water has two main absorption bands at 3546 A (c = 23)
and 2100 A (e = 5380) and a weak band at 2870 A (v = 9). The assignment of
these bands has been reported,(1 9) and this article is an excellent reference for
gaining an appreciation of how the concepts discussed in this chapter are used
1
in band assignments. The band at 3546 A is an n -* nr* transition ( B, <- 'A 1 )
involving the oxygen lone pair. The band at 2100 A is assigned as
E -+ 2*('B 2 - 'A1), and the band at 2870 A is assigned to an n --* r* transition
('A 2 +- 'A) involving the oxygen lone pair.
The absorption peaks obtained for various inorganic anions in water or
alcohol solution are listed in Table 5-3. For the simple ions (Br-, Cl-, OH-)
the absorption is attributed to charge transfer in which the electron is transferred
to the solvent.
There are many more examples of applications of electronic spectroscopy
to inorganic and organometallic systems. These are reviewed on a regular basis
in the Specialist Periodical Reports of the Chemical Society (London). When well
armed with the fundamentals, this source(2 0 >provides more examples.
TABLE 5-3. Characteristic Absorption

Maxima for Some Inorganic Anions
Cl- 1810 104

Br- 1995 11,000
1900 12,000
I 2260 12,600
1940 12,600
OH- 1870 5000
SH- 2300 8000
S2O32- 2200 4000
S20 2 2540 22
NO 2 3546 23
2100 5380
2870 9
NO 3 3025 7
1936 8800
N2O22- 2480 4000
5-17 Molecular Addition Compounds of Iodine 133
5-17 MOLECULAR ADDITION COMPOUNDS OF IODINE
The absorption band maximum for iodine (core plus ah 2 4 4 4

r n 1r* ) occurs at about
5200 A in the solvent CCl4 and is assigned to a 7r* -- a* transition. When a donor
molecule is added to the above solution, two pronounced changes in the spectrum
occur (see Fig. 5-11).
C3)
')4)
C
0()
(2)
2 (1, 3
250 350 450 490 520

Wavelength (my)
FIGURE 5-11 Spectra of iodine and base-iodine solutions.

(1) 12 in CC14 ; (2, 3, 4) same 12concentration but increasing
base concentration in the order 2 < 3 < 4.
A blue shift is detected in the iodine peak, and a new peak arises in the
ultraviolet region that is due to a charge-transfer transition.12" The existence of
an isosbestic point at 490 mp indicates that there are only two absorbing species
in the system; namely, free iodine and the complex B: I-II. As indicated in
Chapter 4, a 1:1 equilibrium constant can be calculated from absorbance
measurements for this system. The constant value for K obtained over a wide
range of donor concentrations is evidence for the existence of a 1:1 adduct.
The bonding in iodine adducts can be described by the equation:
= afcrv + ble
where 0e includes contributions from purely electrostatic forces while 0c.,

includes contributions from covalent interactions (these are described as charge-
transfer interactions). In most adducts b > a in the ground state and the band
around 2500 A arises from a charge-transfer transition in which an electron from
this ground state is promoted to an excited state in which a > b. In view of these
coefficients, the charge-transfer band assignment can be approximated by a
transfer of a base electron, nb, to the iodine 6* orbital. These facts and the blue
shift that occurs in the normal 7r* --+u* iodine transition upon complexation can
be explained by consideration of the relative energies of the molecular orbitals
of iodine and the complex (Fig. 5-12). In Fig. 5-12, n, refers to the donor orbital
on the base, and a,2* and 7r2* refer to the free iodine antibonding orbitals involved
in the transition leading to iodine absorption. The oc, 2 *, and ac* are labels for
molecular orbitals in the complex that are very much like the original base and
iodine orbitals because of the weak Lewis acid-base interaction (2 to 10 kcal).
The orbitals nb and o, * combine to form molecular orbitals in the complex, o
and uc*, in which oc, the bonding orbital, is essentially n. and ac* is essentially
E
FIGURE 5-12 Some of the molecular E
orbitals in a base-iodine addition R
compound. G (3)
Y
2 *. Since a,* is slightly higher in energy than the corresponding 6a*, the
transition in complexed iodine [arrow (2) in Fig. 5-12] requires slightly more
energy than the corresponding transition in free 12 [arrow (1)] and a blue shift
is observed. The charge-transfer transition occurs at higher energy in the
ultraviolet region and is designated in Fig. 5-12 by arrow (3).
Some interesting correlations have been reported, which claim that the blue
shift is related to the magnitude of the base-iodine interaction, i.e., the enthalpy
of adduct formation. 2" This would be expected qualitatively from the treatment
in Fig. 5-12 as long as the energy of nr,* differs very little from that of ng* or
else its energy changes in a linear manner with the enthalpy, AH. A rigorous
evaluation of this correlation with accurate data on a wide range of different
types of Lewis bases indicates that a rough general trend exists, but that a
quantitative relation (as good as the accuracy of the data) does not exist. A
relationship involving the charge transfer band, the ionization potential of the
22 23
base, I, and the electron affinity of the acid, Ea, is also reported( , ):
V= Ib-E--A (5-9)
where A is an empirically determined constant for a related series of bases.

The enthalpies for the formation of these charge-transfer complexes are of
interest and significance to both inorganic and organic chemists. For many
inorganic systems, especially in the areas of coordination chemistry and
nonaqueous solvents, information about donor and acceptor interactions is
TABLE 5-4. Equilibrium Constants and Enthalpies of

Formation for Some Donor-12 Adducts
Donor K(liter mole -1) -AH(kcal mole-')
C6 H 6 0.15 (25 ) 1.4

Toluene 0.16 (25 ) 1.8
CH 3 OH 0.47 (200) 1.9
Dioxane 1.14 (17 ) 3.5
(C2 H 5)20 0.97 (20') 4.3
(C2 H5)2 S 180 (25 ) 8.3
CH 3C(O)N(CH 3 )2 6.1 (25 ) 4.7
Pyridine 270 (200) 7.8
(C2H5)3 N 5130 (250) 12.0
5-18 Effect of Solvent Polarity on Charge Transfer Spectra 135
essential to an understanding of many phenomena. Since the above adducts are
soluble in CCl4 or hexane, the thermodynamic data can be interpreted more
readily than results obtained in polar solvents, where large solvation enthalpies
and entropies are encountered. Some typical results from donor-12 systems in
which such solvation effects are minimal illustrate the wide range of systems that
can be studied, and are contained in Table 5-4.
The following few examples illustrate the information that can be obtained
by studying enthalpies of association in non-polar, weakly basic solvents.
1. The donor properties of the 71 electron systems of alkyl-substituted
benzenes have been reported." 4
2. A correlation of the heat of formation of iodine adducts of a series of
para-substituted benzamides with the Hammett substituent constants' 2 5 , of the
benzamides is reported.
3. The donor properties of a series of carbonyl compounds [(CH 3 )2 CO,
CH 3C(O)N(CH 3)2, (CH 3)2 NC(O)N(CH3)2, CH 3C(O)OCH 3, CH 3C(O)SCH 3]
have been evaluated and interpreted12 1> in terms of conjugative and inductive
effects of the group attached to the carbonyl functional group.
4. The donor properties of sulfoxides, sulfones, and sulfites have been
investigated.1 2 The results are interpreted to indicate that sulfur-oxygen 7
bonding is less effective in these systems than carbon-oxygen 7a bonding is in
ketones and acetates.
5. The effect of ring size on the donor properties of cyclic ethers and sulfides
has been investigated.(2 8 ) It was found that for saturated cyclic sulfides, of general
formula (CH2)n S, the donor properties of sulfur are in the order n = 5 > 6 > 4 > 3.
The order for the analogous ether compounds is 4 > 5 > 6 > 3. Explanations of
these effects are offered.
6. The donor properties of a series of primary, secondary, and tertiary
amines have been evaluated.129 30 ' ) The order of donor strength of amines varies
with the acid studied. Explanations have been offered, which are based upon the
relative importance of covalent and electrostatic contributions to the bonding in
various adducts.
In addition to iodine, several other Lewis acids form charge-transfer
complexes that absorb in the ultraviolet or visible regions. For example, the
relative acidities of 12, ICI, Br2, SO 2 , and phenol toward the donor N,N-
dimethylacetamide have been evaluated. Factors affecting the magnitude of the
interaction 3 1 ) and information regarding the bonding in the adducts are reported.
9 3 2 3 3 34 5
Good general reviews of charge-transfer complexes are available.! , . , .3 )
5-18 EFFECT OF SOLVENT POLARITY ON CHARGE-

TRANSFER SPECTRA
The ion pair N-methylpyridinium iodide undergoes a charge-transfer transition

that can be represented( 36 ) as in Fig. 5-13. It has been found that the position
of the charge-transfer band is a function of the solvating ability of the solvent.
A shift to lower wavelengths is detected in the better solvating solvents. The
positions of the bands are reported as transition energies, ET. Transition energies
(kcal mole') are calculated from the frequency as described in Chapter 4. The
transition energy is referred to as the Z-value. Some typical data are reported
FIGURE 5-13 Ion pair and charge 'e-

transfer excited state of N N
N-methylpyridinium iodide. CH 3
Ion pair Excited state
3
in Table 5-5. An explanation for tne observed shift has been proposed.! ,a) The
dipole moment of the ion pair, C 5H 5NCH 3 +I , is reported to be perpendicular
to the dipole moment of the excited state (Fig. 5-13). Polar solvent molecules
will align their dipole moments for maximum interaction with the ground state,
lowering the energy of the ground state by solvation. The dipole moment of the
solvent molecules will be perpendicular to the dipole moment of the excited state,
producing a higher energy for the excited state than would be found in the gas
phase. Since solvent molecules cannot rearrange in the time required for a
transition, the relative lowering of the ground state and raising of the excited
state increases the energy of the transition, E,, over that in the gas phase (Fig.
5-14), shifting the wavelength of absorption to higher frequencies. Hydrogen-
bonding solvents are often found to increase ET more than would be expected
by comparing their dielectric constants with those of other solvents. This is due
to the formation of hydrogen bonds with the solute. The use of the dielectric
constant to infer solvating ability can lead to difficulty because the local dielectric
constant in the vicinity of the ion may be very different from the bulk dielectric
constant.
I
Excited
Excited state
FIGURE 5-14 Effect of solvent on the

state ET
E I
transition energy, Er-
TABLE 5-5. Ion pair
Z-Values for Some Ion pair
Common Solvents Gas phase Solution
Solvent ET or Z value'
H2 0 94.6
CH 30H 83.6 The data obtained from these spectral shifts are employed as an empirical
C2HOH 79.6 measure of the ionizing power of the solvent. The results can be correlated with
CH 3 COCH, 65.7
(CH 3) 2 NCHO 68.5
a scale of "solvent polarities" determined from the effect of solvent on the rate
CH 3CN 71.3 of solvolysis of t-butyl chloride.!3 a) Other applications of these data to kinetic
Pyridine 64.0 and spectral studies are reported.(3 sb> Solvent effects are quite complicated, and
CHSOCH3 71.1 these correlations at best provide a semiquantitative indication of the trends
H2 NCHO 83.3
CH 2 Cl2 64.2 expected.
Isoctane 60.1 Significant differences exist between "solvating power" inferred from the
dielectric constant and the results from spectral and kinetic parameters. Although
' The ET or Z value is the transition
energy in kcal mole- at 25 C, 1
methanol and formamide are found to have similar Z-values, the dielectric
atm pressure, for the compound 1- constants are 32.6 and 109.5, respectively. Solvent effects cannot be understood
ethyl -4 - carbomethoxypyridinium solely on the basis of the dielectric constant. Specific Lewis acid-base interactions
iodide. make the problem more complex than the simple dielectric model.
The band positions for the n - it* transitions in certain ketones in various
solvents are found to be linearly related to the Z-values for the solvents. A
constant slope is obtained for a plot of ET versus Z for many ketones. Deviations
from linearity by certain ketones in this plot can be employed to provide
interesting structural information about the molecular conformation. Cyclo-
heptanone, for example, does not give a linear plot of ET versus Z. The deviation
is attributed to solvent effects on the relative proportion of the conformers present
in solution.
5-19 STRUCTURES OF EXCITED STATES
Considerable information is available about the structure of excited states of

molecules from analysis of the rotational band contours in the electronic spectra.
Both geometrical information and vibrational information about the excited
states of large molecules can be obtained."" By studying and analyzing the
perturbation made on the vibrational fine structure of an electronic transition
by an electric field, the dipole moment of the excited state can be obtained.(3 8 )
5-20 INTRODUCTION Optical Rotary

Plane-polarized light consists of two circularly polarized components of equal
Dispersion,
intensity. The two types of circularly polarized light correspond to right-handed Circular
and left-handed springs. Circularly polarized light is defined as right-handed
when its electric or magnetic vector rotates clockwise as viewed by an observer
Dichroism, and
facing the direction of the light propagation (i.e., the source). The frequency of Magnetocircular
the rotation is related to the frequency of the light. Plane polarized light can be
resolved into its two circular components, and the two components when added
Dichroism
together produce plane-polarized light in an optically isotropic medium. If
plane-polarized light is passed through a sample for which the refractive indices
of the left- and right-polarized components differ, the components will, upon
recombination, give plane-polarized radiation in which the plane of the polari-
zation has been rotated through an angle 2, given by
a=n, - n,'r (5-10)
where the subscripts refer to left and right, n is the appropriate refractive index,
and 2 is the wavelength of light employed. The units are radians per unit length,
with the length units given by those used for A.
If the concentration of an optically active substance, c', is expressed in units
of g cm - (corresponding to the density for a pure substance), the specific rotation
[a] is defined as:
[2] = ' (5-11)

where d' is the thickness of the sample in decimeters. The molar rotation [M]
is defined as:
[M] = M[2c] x 10- 2 = Mo( x 10- 2 /c'd' (5-12)
where M is the molecular weight of the optically active component. (The quantity
10-2 is subject to convention and not always included in [M].)
The optical rotatory dispersion curve, ORD, is a plot of the molar rotation,
[Lx] or [M), against 2. When the plane of polarization rotates clockwise as viewed
by an observer facing the direction of propagation of the radiation, [c] or [M]
is defined as positive; a counterclockwise rotation is defined as negative.
The technique whereby one determines that an optically active substance
absorbs right and left circularly polarized light differently is called circular
dichroism, CD. All optically active substances exhibit CD in the region of
appropriate electronic absorption bands. The molar circular dichroism 8, - E,is
defined as
r, - e, = k, -k (5-13)
c
where k, the absorption coefficient, is defined by I = In 10 -kd with I, and I being

the intensity of the incident and resultant light and d being the cell thickness.
The CD curve results when we plot e, - r, versus ).
Wherever circular dichroism is observed in a sample, the resulting radiation
is not plane polarized, but is elliptically polarized. The quantity c in the above
equations is then the angle between the initial plane of polarization and the major
axis of the ellipse of the resultant light. One can define a quantity q' (in radians),
the tangent of which is the ratio of the major to minor axes of the ellipse. The
quantity cp' (in radians), the tangent of which is the ratio of the major to minor
axes of the ellipse. The quantity p' is used to approximate the ellipticity; when
it is expressed in degrees, it can be converted to a specific ellipticity [q] or molar
ellipticity [0] by
[p] = c(5-14)
c d'
and
[0] = M[(]10 -2 (5-15)
where the symbols are as defined in equations (5-11) and (5-12). The quantity
[0] is related to E,- r,. by the following equation:
r, - e, = 0.3032 x 10 3[Q] (5-16)
Thus, one often sees the CD curve plotted as [0] versus 2.

With CD one can measure only the optical activity if there is an
accompanying electronic absorption band. On the other hand, ORD is measurable
[a] 0 \
- % I
NJ
SI I
300 500
X(my)
FIGURE 5-15 ORD (---) and CD (-) curves for
D-(-). [Rh(en) 3]3+ (en is an abbreviation for
ethylenediamine).
both inside and outside the absorption band.* The ORD and CD curves of
D-(-)-[Rh(en)3]"3are illustrated( 3 9) in Fig. 5-15. Throughout most of the visible
region, the ORD curve is negative. However, the CD curve associated with the
visible d-d transitions at ~300 my is clearly positive. All the chromophores in a
molecule contribute to the rotatory power at a given wavelength, but only the
chromophore that absorbs at the given wavelength contributes to the CD. Thus,
a transition in the far u.v. can make a significant contribution to the rotation in
the region of d-d transitions in the ORD. The negative effect in the u.v. dominates
the d-d contribution through most of the visible region, and the negative ORD
curve results. For most of the applications to be discussed here, CD is the method
of choice.
5-21 SELECTION RULES
We have previously given [equation (5-3)] the transition moment integral for
an electric dipole transition, and we mentioned that a magnetic dipole transition
integral has a similar form. In order for an electronic transition to give rise to
optical activity, the transition must be both electric and magnetic dipole allowed,
i.e.,
R oc If e,$ e,*x dv][f feiMDO ex dv]
where R is the rotational strength, MD is the magnetic dipole operator, and M

is the electric dipole operator. If an electronic absorption band is observed, there
must be some mechanism for making this allowed; it then becomes important
* CD is the absorption difference between left and right circularly polarized light for an electronic
transition. ORD is related to the difference between the indices of refraction for left and right circularly
46
polarized light. The two effects are interrelated via the Kronig-Kramers relation.
to be concerned with the magnetic dipole selection rules. The sign and magnitude
of the activity can be calculated(")0 by evaluating both the electric and magnetic
dipole integrals.
5-22 APPLICATIONS
The use of CD in band assignments is an obvious application of our previous

discussion of the selection rules. For example, octahedral nickel(II) complexes
have three bands assigned as 3A2, to 3T, 3 T,(F), and 3Tg(P) in order of
increasing energy. Since the magnetic dipole operator (the magnetic dipole
3
transforms like the rotations RRYR,) in 0 , is Tig, only the 22, -* T2g transition
is magnetic dipole allowed. Accordingly, it is found in the CD spectrum of
Ni(pn) 32 + (pn = propylene diamine) that the low-energy band has a maximum
(e, - r,) value of 0.8, while those for the other bands are less than 0.04. This
confirms the original band assignments.
A second type of application involves using CD to show that certain
absorption bands have contributions from more than one electronic transition.
The CD bands are usually narrower and can be positive or negative. This idea
is illustrated in Fig. 5-16, where the absorption curve and CD curve for
A(+)-Co(en) 33+ in aqueous solution are shown.(4 1 )
Co(III) complexes with 0 h symmetry commonly have two absorptions
assigned to 'Ti, and 'T2g. The symmetry of Co(en) 3 3+ is D3, so the transitions
to the T states are expected to show some splittings. This is not detected in the
absorption spectrum, as seen in Fig. 5-16. However, the effects of lower symmetry
are observed in the CD spectrum. The magnetic dipole selection rules for D3
predict that the low-energy band will have two magnetic dipole allowed
components, 'A, - 'A 2 and 'A, -+ 'E. The 'A, -* 'E, transition of the high-
energy band is magnetic dipole allowed, but the 'A, -+ 'A, transition is not. In
the CD, two components (+ and -) are seen in the low-energy band and one
in the high-energy band, as predicted for a D3 distortion. Applications of these
ideas can be used to indicate the symmetry of molecules in solution and in uniaxial
single crystals.
FIGURE 5-16 The

absorption spectrum (solid 2
line) and CD curve (dashed
3 1
line) of (+)-Co(en) 3 + in
aqueous solution. ' 0
-- 1
cm-I
5-23 Magnetocircular Dichroism 141
Another application involves using the sign of the CD to obtain the absolute
configuration of a molecule.(4" This application has been particularly successful
for organic compounds.( 413 In organic systems, absolute configurations are often
assigned by analogy to known systems. Particular care must be employed in
determining what constitutes an analogous compound.(4 4 4 By using complete
operator-matrices for the electric and magnetic dipole components, the signs of
the trigonal components in some Co(III) and Cr(III) complexes have been related
to the absolute configurations of the complexes.(40 ).
Finally, optically active transitions are polarized, and the polarization
information can be used to support the assignment of the electronic spectrum.
5-23 MAGNETOCIRCULAR DICHROISM
When plane-polarized light is passed through any substance in a magnetic field,

HO, whose component in the direction of the light propagation is non-zero, the
substance appears to be optically active. Left and right circularly polarized light
do not interact in equivalent ways. For atoms, for example, left circularly polarized
(lcp) light induces a transition in which Ami is -1, while for right circularly
polarized (rcp) light Amj is + 1. Note that mj has the same relationship to J as
m, has to 1. If one observed a transition from an S state where J = 0 to a P
state where J = 1, the two transitions in Fig. 5-17 would occur as a consequence
of this selection rule.(4 6 )
If the absorptions of left and right circularly polarized light corresponding
to these transitions are measured separately, the curves in Fig. 5-17(B) are
obtained. Here vo is the band maximum for the absorption band. When the mcd
curve is plotted, the result in Fig. 5-17(C) is obtained, provided that the band
width is much greater than the Zeeman splitting of the excited state. A curve of
this sort is referred to as an A-term and can arise only for a transition in which
J > 0 for one of the states involved. The sign of the A-term in molecules depends
upon the sign of the Zeeman splitting and the molecular selection rules for
circularly polarized light.
1p - +1
-- 0
0 CQ +
M +
rcp
HoV V0
HO
er
(A) (B) (C)

FIGURE 5-17 The transitions and expected spectrum for 'S-+'P in the mcd
experiment. (A) the transitions; (B) spectra for left (s,) and right (c,) circularly
polarized radiation; (C) the mcd spectrum (E,- e,) circularly polarized radiation; (C)
the mcd spectrum (e, - E,); A-term behavior.
Next consider a transition from a P to a 'S state. Figure 5-18 summarizes

this situation. The Am, = + 1 transition for rcp light is now that of m. from - 1
to 0. Because the m. states are not equally populated, but Boltzmann populated,
the two transitions will not have equal intensity, as shown in Fig. 5-18B. The
relative intensities will be very much temperature dependent. The resultant mcd
curve, shown in Fig. 5-18C, is referred to as a C-term. The band shape and
intensity are very temperature dependent. An A-term curve usually occurs
superimposed upon a C-term curve.
is0
lcp rcp
1P-
--- 1 er
FIGURE 5-18 The transitions and expected spectrum for IP-f 'S in the mcd
experiment. (A) the transitions; (B) spectra for left and right circularly polarized
radiation; (C) the mcd spectrum; C-term behavior.
A third type of curve (B-term) results when there is a field-induced mixing

of the states involved (this phenomenon also creates temperature independent
paramagnetism, TIP, and will be discussed in more detail in the chapter on
magnetism). This is manifested in a curve that looks like a C-curve but that is
temperature independent. Since this mixing is present to some extent in all
molecules, all substanceshave mcd activity. The magnitude of the external magnetic
field intensity will determine whether or not the signal is observed.
The following characteristics summarize the basis for detecting and
qualitatively interpreting mcd curves:
1. An A-term curve changes sign at the absorption maximum, while B-
and C- curves maximize or minimize at the maximum of the electronic absorption
band.
2. A C-term curve's intensity is inversely proportional to the absolute
temperature, while a B-term is independent of temperature.
3. An A-term spectrum is possible only if the ground or excited state
involved in the electronic transition is degenerate and has angular momentum.
4. A C-term spectrum is possible only if the ground state is degenerate and
has angular momentum.
4 6 4 7
As can be anticipated, mcd measurements are of considerable utility( , )
in assigning the electronic spectrum of a compound. Furthermore, the magnitude
of the parameters provides information about many subtle electronic effects.(47)
The molecular orbital origin of an electron involved in a transition can be
determined. The lowest energy band in RuO 4 148 is clearly an oxygen to ruthenium
charge-transfer band. One cannot determine from the electronic spectrum whether
the oxygen electron involved in the transition came from a tix or t 2 7r type of
oxygen molecular orbital. The sign of the mcd A-term established( 481 the transition
as t 11(oxygen) -* eg - (ruthenium). Another significant advantage of mcd is
in the assignment of spin-forbidden electronic transitions that have very low
intensity in the electronic absorption spectrum. The assignments of the
components in a six-coordinate chromium(III) complex have been made with
this technique.(4 9 ) Other applications have been summarized in review
articles.(4 6 471 ) Mcd has been extensively applied( 46 b) to provide information
regarding the symmetries, angular momenta, electronic splittings, and
vibrational-electronic interactions in excited electronic states.
1. H. H. Jaffe and M. Orchin, "Theory and Applications of Ultraviolet Spectroscopy," REFERENCES

Wiley, New York, 1962. CITED
2. M. Kasha, Discussions Faraday Soc., 9, 14 (1950).
3. M. Kasha, Chem. Revs., 41, 401 (1947); J. R. Platt, J. Opt. Soc. Amer., 43, 252 (1953):
A. Terenin, Acta Physica Chim. USSR, 18, 210 (1943) (in English).
4. M. Kasha, "The Nature and Significance of n - 7r* Transitions," in Light and Life,
ed. by W. D. McElroy and B. Glass, Johns Hopkins, Baltimore, 1961.
5. J. Sidman, Chem. Revs., 58, 689 (1958),.
6. H. McConnell, J. Chem. Phys., 20, 700 (1952).
7. G. J. Brealey and M. Kasha, J. Amer. Chem. Soc., 77, 4462 (1955).
8. S. Nagakura, Bull. Chem. Soc. Japan, 25, 164 (1952) (in English).
9. L. E. Orgel, Quart. Revs., 8, 422 (1954). For a good discussion of charge-transfer
transitions, see A. B. Lever, J. Chem. Ed., 51, 612 (1974).
10. S. P. McGlyn and M. Kasha, J. Chem. Phys., 24, 481 (1956).
11. C. K. Jorgenson, Acta Chem. Scand., 11, 166 (1957); R. J. P. Williams, J. Chem. Soc.,
1955, 137.
12. R. P. Bauman, "Absorption Spectroscopy," Wiley, New York, 1962.
13. L. Lang, "Absorption Spectra in the Ultraviolet and Visible Region," Vols. 1-20,
Academic, New York, 1961-76.
14. H. M. Hershenson, "Ultraviolet and Visible Absorption Spectra-Index for 1930-54,"
Academic, New York, 1956.
15. "Organic Electronic Spectral Data," Vols. 1-10, Interscience, New York. Compilation
of spectral data from 1946 to 1968.
16. "ASTM (American Society for Testing Materials) Coded IBM Cards for Ultraviolet
and Visible Spectra" ASTM, Philadelphia, 1961.
17. S. F. Mason, Quart. Revs., 15, 287 (1961).
18. N. Metropolis, Phys. Rev., 60, 283, 295 (1941).
19. S. J. Strickler and M. Kasha, J. Amer. Chem. Soc., 85, 2899 (1963).
20. "Spectroscopic Properties of Inorganic and Organometallic Compounds." Volumes
1- (1967- ), Specialist Periodical Report, Chemical Society, London.
21. R. Foster, "Organic Charge Transfer Complexes," Academic Press, London, 1969.
22. H. McConnell, J. S. Ham and J. R. Platt, J. Chem. Phys., 21, 66 (1953).
23. G. Briegleb and J. Czekalla, Z. Phys. Chem. (Frankfurt), 24, 37 (1960).
24. R. M. Keefer and L. J. Andrews, J. Amer. Chem. Soc., 77, 2164 (1955).
25. R. L. Carlson and R. S. Drago, J. Amer. Chem. Soc., 85, 505 (1963).
26. R. L. Middaugh, R. S. Drago and R. J. Niedzielski, J. Amer. Chem. Soc., 86, 388 (1964).
27. R. S. Drago, B. Wayland and R. L. Carlson, J. Amer. Chem. Soc., 85, 3125 (1963).
28. Sister M. Brandon, 0. P., M. Tamres and S. Searles, Jr., J. Amer. Chem. Soc., 82,
2129 (1960); M. Tamres and S. Searles, Jr., J. Phys. Chem., 66, 1099 (1962).
29. H. Yada, J. Tanaka and S. Nagakura, Bull. Chem. Soc. Japan, 33, 1660 (1960).
30. R. S. Drago, D. W. Meek, R. Longhi and M. Joesten, Inorg. Chem., 2, 1056 (1963).
31. R. S. Drago and D. A. Wenz, J. Amer. Chem. Soc., 84, 526 (1962). Other donors
studied toward these acids are summarized here.
32. L. J. Andrews and R. M. Keefer, "Advances in Inorganic Chemistry and Radiochem-
istry," Vol. 3, eds. H. J. Emeleus and A. G. Sharpe, pp. 91-128, Academic, New York,
1961.
33. 0. Hassel and Chr. Romming, Quart. Revs., 16, 1 (1962).
34. R. Foster, "Molecular Complexes," Crane, Russak and Co., New York, 1974.
35. G. Briegleb, "Electronen Donator-Acceptor-Komplexes," Springer-Verlag, Berlin,
1963.
36. a. E. M. Kosower, et al., J. Amer. Chem. Soc., 83, 3142, 3147 (1961); E. M. Kosower,
J. Amer. Chem. Soc., 80, 3253, 3261, 3267 (1958); E. M.Kosower, "Change Transfer
Complexes," in "The Enzymes," Vol. 3, eds. Boyer, Lardy, and Myrbach, p. 171,
Academic Press, New York, 1960. The Spectrum of N-methylpyridinium iodide in
many solvents is treated exhaustively in these references.
b. C. Reichardt, "Solvents and Solvent Effects in Organic Chemistry," VCH Publishers,
New York, 1988.
37. J. M. Hollas, "Molecular Spectroscopy," Vol. 1, pp. 62-112, Specialist Periodical
Reports, Chemical Society, London (1972).
38. D. E. Freeman and W. Klemperer, J. Chem. Phys., 45, 52 (1966).
39. J. P. Mathieu, J. Chim. Phys., 33, 78 (1936).
40. R. S. Evans, A. F. Schreiner, and J. Hauser, Inorg. Chem., 13,2185 (1974) and references
therein.
41. A. J. McCaffery and S. F. Mason, Mol. Phys., 6, 359 (1963).
42. E. J. Corey and J. C. Bailar, Jr., J. Amer. Chem. Soc., 81, 2620 (1959).
43. L. Velluz, M. Legrand and M. Grosjean, "Optical Circular Dichroism," Academic
44. F. Woldbye, "Technique of Inorganic Chemistry," Vol. 4, eds. H. B. Jonassen and A.
Weissberger, p. 249, Interscience, New York, 1965.
45. A. M. Sargeson, Transition Metal Chemistry, 3, 303 (1966).
46. (a) A. D. Buckingham and P. J. Stephens, Ann. Rev. Phys. Chem., 17, 399 (1966); (b)
P. J. Stephens, Ann. Rev. Phys. Chem., 25, 201 (1974).
47. P. N. Schatz and A. J. McCaffery, Quart. Revs., 23, 552 (1969), and references therein.
48. A. H. Bowman, R. S. Evans and A. F. Schreiner, Chem. Phys. Letters, 29, 140 (1974),
and references therein.
49. P. J. Hauser, A. F. Schreiner and R. S. Evans, Inorg. Chem., 13, 1925 (1974), and
references therein.
ADDITIONAL A. The following references may be consulted for further study:

REFERENCES
H. H. Jaffe and M. Orchin, "Theory and Applications of Ultraviolet Spectroscopy," Wiley,
New York, 1962.
M. Orchin and H. H. Jaffe, "Symmetry, Orbitals and Spectra," Wiley Interscience, New
York, 1971.
C. Sandorfy, "Electronic Spectra and Quantum Chemistry," Prentice-Hall, Englewood
Cliffs, N. J., 1964.
R. P. Bauman, "Absorption Spectroscopy," Wiley, New York, 1962.
E. A. Braude and F. C. Nachod, "Determination of Organic Structures by Physical
Methods," pp. 131-195, Academic, New York, 1955.
G. Herzberg, "Molecular Spectra and Molecular Structure," Vols. 1 and 3. D. Van
Nostrand Co., Princeton, N.J., 1966.
Exercises 145
J. N. Murrell, "The Theory of Electronic Spectra of Organic Compounds," Wiley, New
York, 1963.
A. E. Gillam and E. S. Stern, "Introduction to Electronic Absorption Spectroscopy," 2nd
ed., Arnold, London, 1957.
G. W. King, "Spectroscopy and Molecular Structure," Holt, Rinehart, and Winston, New
York, 1964.
A. B. P. Lever, "Inorganic Electronic Spectroscopy," 2nd ed. Elsevier, New York, 1984.
A. P. Demchenko, "Ultraviolet Spectroscopy of Proteins," Springer-Verlag, Berlin, 1986.
"Spectroscopy in Biochemistry" ed. J. E. Bell, CRC Press, Boca Raton, Fla., 1980.
1. The compound (CH,)3 As is reported [J. Mol. Spectr., 5, 118-132 (1960)] to have EXERCISES
two absorption bands, one near 2700 A and a second around 2300 A. One band is
due to the 7r -- n* transition of the phenyl ring and the other is a charge-transfer
transition from the lone pair electrons on arsenic to the ring. The 2300 band is solvent
dependent, and the 2700 band is not.
a. Which band do you suspect is 7T -> *? -
b. What effect would substitution of one of the phenyl groups by CF 3 have on the
frequency of the charge-transfer transition?
c. In which of the following compounds would the charge-transfer band occur at

highest frequency, (CH)3P, (CH4)3Sb, or (C6 H 5 )3 Bi?
2. What effect would changing the solvent from a nonpolar to a polar one have on the
frequency for the following:
a. Both ground and excited states are neutral (i.e., there is no charge separation)?
b. The ground state is neutral and the excited state is polar?
c. The ground state is polar, the excited state has greater charge separation. and
the dipole moment vector in the excited state is perpendicular to the ground state
moment?
d. The ground state is polar and the excited state is neutral?
3. Would you have a better chance of detecting vibrational fine structure in an electronic
transition of a solute in liquid CCI4 or CH 3CN solution? Why?
4. Do the excited states that result from n - n* and 7T -> i* transitions in pyridine belong
to the same irreducible representation? To which species do they belong?
5. Under what conditions can electronic transitions occur in the infrared spectrum?
Which compounds would you examine to find an example of-this?
6. Explain why the transition that occurs in the ion pair N-methylpyridinium iodide
does not occur in the solvent-separated ion pair.
7. Recall the center of gravity rule and explain why the blue shift in the iodine transition
should be related to the heat of interaction of iodine with a donor (see Fig. 5-12).
8. What is the polarization expected in the 'A,1 'A2(n 7t*) transition of formaldehyde
from vibronic coupling? The vibrational modes have a,, b, and b2 symmetry.
9. Two observed transitions in octahedral cobalt(II) complexes occur at 20,000 cm-

and 8000 cm -1 with extinction coefficients of e - 50 and 8, respectively. These
4
have been assigned to 'Tg-- Tg and 'T,, - T,. Indicate whether or not
vibronic coupling can account for the intensity difference. (The vibrations are ag,
e., t,., t2., and t 2 .)-
2
10. The vibrations of CoCl4 - have symmetries corresponding to a, e, and 2t 2 . Explain
whether or not the A 2 - T2 and A 2 -- T, d-d transitions can gain intensity by
vibronic coupling.
11. Consider the cyclopropene cation shown below:
C C
H H
a. Using the three out-of-plane p orbitals (labeled qP,, (P 2 , and (3) as a basis set, do
a simple Huckel calculation to obtain the energies of the resultant m.o.'s in terms
of a and #l.
b. The symmetries of these m.o.'s in the Ds, point group are A2 " and E". Briefly
describe two methods you could use to find O(A 2") in terms of P1, P2, and (P3 -
c. Show whether the electronic transition to the first excited state is allowed. If so,
what is its polarization?
d. What is the energy of this transition (in terms of a and #l)?
e. What complications are likely in the approach used in part d?
12. Pyrazine has D2 symmetry. The six pi-levels are shown below with a rough order of
energies. The two unlabeled levels are the non-bonding m.o.'s from the nitrogen lone
pairs.
b2 ,
uON NO z
b,.
y
x is out of plane
b2 ,
big
bs,
a. What are the symmetries of the two non-bonding m.o.'s?
b. What is the symmetry of the state of lowest energy that arises from a 7T ->*
transition? Is it allowed? If so, what is its polarization?

Exercises 147
c. What is the symmetry of the states that arise from n - n* transitions from the two
non-bonding levels to the lowest 7t* level? Determine whether either is allowed, and,
if so, give the polarization. If either is forbidden, give the symmetry that a vibrational
mode would need in order to lend intensity via the vibronic mechanism.
13. In the Re 2 Cl 2 - ion, the transition of an electron from the bg orbital to the b, orbital
is a d-d transition in a molecule with a center of inversion. Is it allowed? Explain.
1
- (d.,(" + d,(2>
1
- (d,(') - d,)
/2
14. In several coordination complexes of cis-butadiene, the 2,3 carbon-carbon bond
distance was observed to be appreciably shorter than the normal single bond distance.
The possibility of back donation of electron density from the metal into the it* orbital
of the butadiene was proposed to account for this shortening. This exercise is designed
to provide some insights into the chemistry of this ligand.
a. Do a simple Huckel calculation on the 7t orbitals of cis-butadiene. Obtain the

energies of the 7t-symmetry orbitals. (Hint: set y = x 2.)
b. Work out the symmetries of the it and it* orbitals.
c. The coefficients of the it molecular orbitals are given as follows.

C, C2 C' C4
.60 -. 37 -. 37 .60
b .37 .60 .60 .37

, .37 -. 60 .60 -. 37
'd .60 .37 -. 37 -. 60
(1) Assign the symmetry species of each molecular orbital.

(2) Arrange them in order of increasing energy.
d. Determine the symmetries of all four singly excited states of cis-butadiene. Which
transitions are allowed? Give their polarizations.
e. The lowest-energy it -r* transition is observed at . - 217 my in cis-butadiene.

Calculate #l.
f. Calculate the 2,3 bond order for cis-butadiene and then calculate the 2,3 bond
order with two additional electrons in the lowest it* orbital. Do your results agree
with the conclusions regarding it back bonding?
15. For trans-butadiene,
a. determine the symmetries of the it molecular orbitals.
b. determine the symmetries of all four singly excited states. Which transitions are
allowed? Give their polarizations.
c. compare the results in parts a and b with those obtained for cis-butadiene. Would
an HMO calculation differentiate between these two rotamers?
16. Consider the 7c molecular orbitals obtained from the p. atomic orbitals in NO 3 . The
ground state orbital energies and symmetries are:
a2
e
a2"
a. Obtain the state symmetries of the singly excited states.
b. Which electronic transitions are allowed? Explain.
c. NO, -has four vibrational modes, A, 2E', and A2". Does vibronic coupling provide
a mechanism by which any forbidden transitions become allowed? Explain.
17. Show that the following two statements are equivalent for the point groups D4 , C 3h,
and T:
a. In order to observe circular dichroism, the transition from ground state to excited
state must be simultaneously electric dipole-allowed and magnetic dipole-allowed.
(The magnetic dipole moment along the a axis transforms as a rotation about the
a axis, R,.)
b. In order to observe circular dichroism, the molecule must be optically active.
Vibration and Rotation
Spectroscopy: Infrared,
Raman, and Microwave
6
6-1 HARMONIC AND ANHARMONIC VIBRATIONS Introduction
As discussed earlier (Chapter 4), quanta of radiation in the infrared region have
energies comparable to those required for vibrational transitions in molecules.
Let us begin this discussion by considering the classical description of the
vibrational motion of a diatomic molecule. For this purpose it is convenient to
consider the diatomic molecule as two masses, A and B, connected by a spring.
In Fig. 6-1A the equilibrium position is indicated. If a displacement of A and B
is carried out, moving them to A' and B', [as in Fig. 6-1B], there will be a force
acting to return the system to the equilibrium position. If the restoring force
exerted by the spring, f, is proportional to the displacement Ar, i.e.,
f =- k Ar (6-1)
the resultant motion that develops when A' and B' are released and allowed to
oscillate is described as simple harmonic motion. In equation (6-1), the Hooke's
law constant for the spring, k, is called theforce constant for a molecular system
held together by a chemical bond.
For harmonic oscillation of two atoms connected by a bond, the potential
energy, V is given by
2
V =kX
where X is the displacement of the two masses from their equilibrium position.
A plot of the potential energy of the system as a function of the distance x between
the masses is thus a parabola that is symmetrical about the equilibrium
internuclear distance, re, as the minimum (see Fig. 6-2). The force constant. k, is
(B)
a measure of the curvature of the potential well near re.
This classical springlike model does not hold for a molecule because a FIGURE 6-1 Displacement of
molecular system cannot occupy a continuum of energy states, but can occupy the equilibrium position of
two masses connected by a
only discrete, quantized energy levels. A quantum mechanical treatment of the
spring.
molecular system yields the following equation for the permitted energy states
of a molecule that is a simple harmonic oscillator:
E,, = hv(v + 2) (6-2)

150 Chapter 6 Vibration and Rotation Spectroscopy: Infrared, Raman, and Microwave
- ------- -- --
FIGURE 6-2 Potential A

7
energy versus distance, x,
for (A) a harmonic oscillator
(solid line) and (B) an
anharmonic oscillator
(dotted line).
where v is an integer 0, 1, 2, ... , representing the vibrational quantum number

of the various states, E, is the energy of the vth state, h is Planck's constant, and
v is thefundamental vibrationfrequency(sec-')(i.e., the frequency for the transition
from state v = 0 to v = 1). These states are indicated for a harmonic oscillator
in Fig. 6-3.
The potential energy curve of a real molecule (see Fig. 5-1), reproduced as
a dotted line in Fig. 6-2, is not a perfect parabola. The vibrational energy levels
are indicated in Fig. 5-1; they are not equally spaced, as equation (6-2) required,
but converge. The levels converge because the molecule undergoes anharmonic
131 rather than harmonic oscillation, i.e., at large displacements the restoring force
(1) 121 'is less than predicted by equation (6-1). Note that as the molecule approaches
dissociation, the bond becomes easier to stretch than the harmonic oscillator
function would predict. This deviation from harmonic oscillation occurs in all
FIGURE 6-3 Vibrational molecules and becomes greater as the vibrational quantum number increases. As
state corresponding to a will be seen later, the assumption of harmonic oscillation will be sufficiently
normal vibrational mode in a accurate for certain purposes (e.g., the description of fundamental vibrations) and
harmonic oscillator, is introduced here for this reason.
6-2 ABSORPTION OF RADIATION BY MOLECULAR

VIBRATIONS-SELECTION RULES
The interaction of electromagnetic, infrared radiation with a molecule involves

interaction of the oscillating electric field component of the radiation with an
oscillating electric dipole moment in the molecule. Thus, in orderfor molecules
to absorb infrared radiationas vibrationalexcitation energy, there must be a change
in the dipole moment of the molecule as it vibrates. Consequently, the stretching
of homonuclear diatomic molecules will not give rise to infrared absorptions.
According to this selection rule, any change in direction or magnitude of the
dipole during a vibration gives rise to an oscillating dipole that can interact with
the oscillating electric field component of infrared radiation, giving rise to
absorption of radiation. A vibration that results in a change in direction of the
dipole is illustrated by the N-C-H bending mode of HCN. There is little change
in the magnitude of the dipole, but an appreciable change in direction occurs
when the molecule bends.
The second selection rule can be derived from the harmonic oscillator
approximation. This selection rule, which is rigorous for a harmonic oscillator,
6-3 Force Constant 151
states that in the absorption of radiation only transitions for which Av = + 1

can occur. Since most molecules are in the v, vibrational level at room
temperature, most transitions will occur from the state vo to v. The transition
is indicated by arrow (1) of Fig. 6-3. The frequency corresponding to this energy
is called the fundamental frequency. According to this selection rule, radiation
with energy corresponding to transitions indicated by arrows (2) and (3) in Fig.
6-3 will not induce transitions in the molecule. Since most molecules are not
perfect harmonic oscillators, this selection rule breaks down and transitions
corresponding to (2) and (3) do occur. The transition designated as (2) occurs at
a frequency about twice that of the fundamental (1), whereas (3) occurs at a
frequency about three times that of the fundamental. Transitions (2) and (3) are
referred to as the first and second overtones, respectively. The intensity of the first
overtone is often an order of magnitude less than that of the fundamental, and
that of the second overtone is an order of magnitude less than the first overtone.
6-3 FORCE CONSTANT
The difference in energy, AE, between two adjacent levels, E. and E,,, is given
by equation (6-3) for a harmonic oscillator:
AE = (6-3)
where k is the stretching force constant and y is the reduced mass

[u mAmB/(mA + mB) for the diatomic molecule A-B]. The relationship between
energy and frequency, AE = hv = hcv, was presented in Chapter 4. The symbol
v will be used interchangeably for frequency (sec-) or wavenumber (cm-'), but
the units will be indicated when necessary. In the HCI molecule, the absorption
of infrared ratiation with v = 2890 cm-' corresponds to a transition from the
ground state to the first excited vibrational state. This excited state corresponds
to a greater amplitude and frequency for the stretching of the H-Cl bond.
Converting v to energy produces AE of equation (6-3). Since all other quantities
are known, this equation can be solved to produce: k = 4.84 x 10' dynes cm-'
or, in other commonly used units, 4.84 md A -. In this calculation the reduced
mass is expressed in grams. Stretching force constants for various diatomic
molecules are summarized in Table 6-1.
The force constants in Table 6-1 are calculated by using equation (6-3), which
was derived from the harmonic oscillator approximation. When an anharmonic
oscillator model is employed, somewhat different values are obtained. For
example, a force constant of 5.157 x 105 dynes cm-' results for HCL. The latter
value is obtained by measuring the first, second, and third overtones and
evaluating the anharmonicity from the deviation of these frequencies from 2, 3,
and 4 times the fundamental, respectively. Since these overtones are often not
detected in larger molecules, we shall not be concerned with the details of the
anharmonicity calculation.
The force constants for some other stretching vibrations of interest are listed
in Table 6-2. For larger molecules, the nature of the vibration that gives rise to
a particular peak in the spectrum is quite complex. Accordingly, one cannot
calculate a force constant for a bond by substituting the "carbonyl frequency,"
TABLE 6-1 Stretching Force Constants

for Various Diatomic Molecules
(Calculated by the Harmonic Oscillator
Approximation)
Molecule v (cm-1) k (dynes cm
HF 3958 8.8 X 10'

HCI 2885 4.8 x 105
HBr 2559 3.8 x 10'
HI 2230 2.9 x 10
F2a 892 4.5 x 105
Cl2a 557 3.2 x 105
Br 20 321 2.4 x 105
I2 213 1.7 x 105
CO 2143 18.7 x 101
NO 1876 15.5 x 105
"Observed by Raman spectroscopy.
for example, of a complex molecule into equation (6-3). This will become clearer
TABLE 6-2. Stretching as we proceed and is mentioned here as a note of caution. The force constants
Force Constants for in Table 6-2 result from a normal coordinate analysis, which will also be discussed
Various Stretching in more detail shortly. A larger force constant is often interpreted as being
Vibrations (Harmonic indicative of a stronger bond, but there is no simple relation between bond
Oscillator Approximation) dissociation energy and force constant. We defined the force constant earlier as a
measure of the curvature of the potential well near the equilibrium internuclear
Bond k (dynes cm -1) configuration. The curvature is the rate of change of the slope, so the force
constant is the second derivative of the potential energy as a function of distance:
C- C:- 4.5 x 10'
C-CE 5.2 x 101
C=CZ 9.6 x 105 k = r2 -> (6-4)
-C C-- 15.6 x 105
_'-C=O 12.1 x 10'
-CN 17.7 x 105 Here V is the potential energy and r is the deviation of the internuclear
-C-H 5.9 x 10 distance, at which r = 0. In a more complicated molecule, r is replaced by q,
-- H 4.8 x 10' which is a composite coordinate that describes the vibration.
Triple bonds have stretching force constants of 13 to 18 x 10', double bonds
about 8 to 12 x 10', and single bonds below 8 x 10' dynes cm -1. In general,
force constants for bending modes are often about a tenth as large as those for
stretching modes.
The bands in the 4000 cm to 600 cm region of the spectrum mostly
involve stretching and bending vibrations. Most of the intense bands above 2900
cm involve hydrogen stretching vibrations for hydrogen bound to a low-mass
atom. This frequency range decreases as the X-H bond becomes weaker and
the atomic weight of X increases. Triple bond stretches occur in the 2000 to 2700
cm-1 region. Absorption bands assigned to double bond stretches occur in the
1500 to 1700 cm -1 region. Bands in this region of the spectrum of an unknown
molecule are a considerable aid in structure determination. Bands in the 1500 to
400 cm-' region find utility in the fingerprint type of application.
Metal-ligand vibrations usually occur below 400 cm-' and into the far
infrared region. They are very hard to assign, since many ligand ring deformation
6-4 The 3N - 6(5) Rule 153
and rocking vibrations as well as lattice modes (vibrations involving the whole
crystal) occur in this region.
As indicated by equation (6-3), the reduced mass is important in determining
the frequency of a vibration. If, for example, a hydrogen bonded to carbon is
replaced by deuterium, there will be a negligible change in the force constant
but an appreciable change in the reduced mass. As indicated by equation (6-3),
the frequency should be lower by a factor of about 1/2. The frequencies for
C-H and C-D vibrations are proportional to [12 x 1/(12 + 1)]- 12 and
[12 x 2/(12 + 2)] - , respectively. Normally, a vibration involving hydrogen will
occur at 1.3 to 1.4 times the frequency of the corresponding vibration in the
deuterated molecule. This is of considerable utility in confirming assignments
that involve a hydrogen atom. The presence of the natural abundance of "C in
a metal carbonyl gives separate bands due to 3 CEO and 2 C--O stretching
vibrations in metal carbonyls (vide infra) because of the difference in reduced
mass. Use of metal isotopes also has utility in confirming the assignment of
vibrations involving the metal-ligand bond."'
6-4 THE 3N- 6(5) RULE Vibrations in a

The positions of the N atoms in a molecule can be described by a set of Cartesian
Polyatomic
coordinates, and the general motion of each atom can be described by utilizing Molecule
three displacement coordinates. The molecule is said, therefore, to have 3N degrees
of freedom. Certain combinations of these individual degrees of freedom correspond
to translational motion of the molecule as a whole without any change in
interatomic dimensions. There are three such combinations which represent the
x, y, and z components of translational motion, respectively. For a nonlinear
molecule there are three combinations that correspond to rotation about the
three principal axes of the molecule without change in interatomic dimensions.
Therefore, for a nonlinear molecule there are 3N - 6 normal modes of vibration
that result in a change in bond lengths or angles in the molecule. Normal modes
represent independent self-repeating motions in a molecule. They correspond to
3N - 6 degrees of freedom that, in the absence of anharmonicity, could take up
energy independently of each other. These modes form the bases for irredicuble
representations. Since a molecule is fundamentally not changed by applying a
symmetry operation R, the normal mode RO must have the same frequency as
the normal mode Q. Thus, if 0 is non-degenerate, RO = ± 1 for all R's.
Consequently, 0 forms the basis for a one-dimensional representation in the
molecular symmetry group. It can be shown that degenerate modes transform
according to irreducible representations of dimensionality greater than one. The
center of mass of the molecule does not change in the vibrations associated with
the normal mode, nor is angular momentum involved in these vibrations. All
general vibrationalmotion that a molecule may undergo can be resolved into either
one or a combination of these normal modes.
For a linear molecule all the vibrations can be resolved into 3N - 5 normal
modes. The additional mode obtained for a linear molecule is indicated in Fig.
6-4(B), where plus signs indicate motion of the atoms into the paper and minus
Center of
rotation
FIGURE 6-4 Rotational and bending

modes for (A)non-linear and (B) linear
molecules.
signs represent motion out of the paper. For the nonlinear molecule [Fig. 6-4(A)]
the motion indicated corresponds to a rotation. For a linear molecule a similar
motion corresponds to a bending of the bonds, and hence this molecule has an
additional normal vibrational mode (3N - 5 for a linear molecule vs. 3N - 6 for
a nonlinear one).
As will be seen later, there are many applications for which we need to know
which bands correspond to the fundamental vibrations.
6-5 EFFECTS GIVING RISE TO ABSORPTION BANDS
Sulfur dioxide is predicted to have three normal modes from the 3N - 6 rule.
The spectral data (Table 6-3) show the presence of more than three bands. The
three bands at 1361, 1151, and 519 cm-' are the fundamentals and are referred
to as the v3, v1, and v 2 bands, respectively (see Fig. 6-5). The v, symbolism is
used to label the various frequencies of fundamental vibrations and should not
be confused with the symbols vU,v 1, V2 , etc., used to designate various vibrational
levels of one mode in a molecule. By convention the highest-frequency totally
symmetric vibration is called v1 , the second highest totally symmetric vibration
v2, etc. When the symmetric vibrations have all been assigned, the highest-
frequency asymmetric vibration is counted next, followed by the remaining
asymmetric vibrations in order of decreasing frequency. An exception is made
to this rule for the bending vibration of a linear molecule, which is labeled v2.
TABLE 6-3. Infrared Another common convention involves labeling stretching vibrations v, bending
Spectrum of S02 vibrations 6, and out-of-plane bending vibrations 7r. Subscripts, as, for asymmetric;
s, for symmetric; and d, for degenerate, are employed with these symbols.
v (cm- 1) Assignment The v, mode in S02 is described as the symmetric stretch, v 3 as the asymmetric
stretch, and v2 as the O-S-0 bending mode. In general, the asymmetric stretch
519 v2 will occur at higher frequency than the symmetric stretch, and stretching modes
606 v1 - V2
1151 occur at much higher frequencies than bending modes. There is a slight angle
v1
1361 V3
change in the stretching vibrations in order for the molecule to retain its center
1871 v2 + v3 of mass. The other absorption frequencies in Table 6-3 are assigned as indicated.
2305 2vi The overtone of v, occurs at about 2v, or 2305 cm-'. The bands at 1871 and
2499 v1 + v3 2499 cm- 'are referred to as combination bands. Absorption of radiation of these
6-5 Effects Giving Rise to Absorption Bands 155
S
T
s
o O / FIGURE 6-5 The three
/ O:
0 0 fundamental vibrations for
k
0sulfur dioxide (The
amplitudes are exaggerated
S S O to illustrate the motion.)
0 0 0 0
v, v3 V2
energies occurs with the simultaneous excitation of both vibrational modes of

the combination. The 606 cm' band is a difference band, which involves a
transition originating from the state in which the v2 mode is excited and changing
to that in which the v, mode is excited. Note that all the bands in the spectrum
are accounted for by these assignments. Making the assignments is seldom this
simple; as will be shown later, much other information, including a normal
coordinate analysis, is required to substantiate these assignments.
A more complicated case is the CO 2 molecule, for which four fundamentals
are predicted by the 3N - 5 rule. A single band results from the two degenerate
vibrations v2 of Fig. 6-6, which correspond to bending modes at right angles to
each other. Later we shall see how symmetry considerations aid in predicting
the number of degenerate bands to be expected. In more complex molecules some
of the fundamentals may be accidentally degenerate because two vibration
frequenciesjust happen to be equal. This is not easily predicted, and the occurrence
of this phenomenon introduces a serious complication. The assignment of the
fundamentals for CO 2 is more difficult than for SO2 because many more bands
appear in the infrared and Raman spectra. Bands at 2349, 1340, and 667 cm-'
have been assigned to v3, v1 , and V2, respectively. The tests of these assignments
have been described in detail by Herzberg (see Additional References) and will
not be repeated here. In this example the fundamentals are the three most intense
bands in the spectrum. In some cases, there is only a small dipole moment change
in a fundamental vibration, and the corresponding absorption band is weak (see
the first selection rule).
The above discussion of the band at 1340 cm - has been simplified. Actually,
it is an intense doublet with band maxima at 1286 and 1388 cm'. This splitting
is due to a phenomenon known as Fermi resonance. The overtone 2v 2
(2 x 667 = 1334 cm- ) and the fundamental v, should occur at almost the same
t1 T
O-c-0
FIGURE 6-6 Carbon dioxide
4 O0->+C-> fundamental vibration modes.
0-C-0
v, - v +2 - v3
frequency. The two vibrations interact by a typical quantum mechanical

resonance, and the frequency of one is raised while the frequency of the other is
lowered. The wave function describing these states corresponds to a mixing of
the wave functions of the two vibrational excited states (vi and 2v2 ) that arise
from the harmonic oscillator approximation. We cannot say that one line
corresponds to vi and the other to 2v2, for both are mixtures of v, and 2v 2 . This
interaction also accounts for the high intensity of what, in the absence of
interaction, would have been a weak overtone (2v 2). The intensity of the
fundamental is distributed between the two bands, for both bands consist partly
of the fundamental vibration.
The presence of Fermi resonance can sometimes be detected in more complex
molecules by examining deuterated molecules or by determining the spectrum
in various solvents. Since the Fermi resonance interaction requires that the
vibrations involved have nearly the same frequency, the interaction will be affected
if one mode undergoes a frequency shift from deuteration or a solvent effect while
the other mode does not. The two frequencies will no longer be equivalent, and
the weak overtone will revert to a weak band or not be observed in the spectrum.
Other requirements for the Fermi resonance interaction will be discussed in the
section on symmetry considerations.
6-6 NORMAL COORDINATE ANALYSES AND BAND

ASSIGNMENTS
Degeneracy, vibration frequencies outside the range of the instruments, low

intensity fundamentals, overtones, combination bands, difference bands, and
Fermi resonance all complicate the assignment of fundamentals. The problem
can sometimes be resolved for simple molecules by a technique known as normal
coordinate analysis. Normal coordinate analysis involves solving the classical
mechanical problem of the vibrating molecule, assuming a particular form of the
potential energy (usually the valence force field). The details of this calculation
are beyond the scope of this text,' 3 1 but it is informative to outline the problem
briefly so the reader can assess the value and the limitations of the approach.
Furthermore, several important qualitative ideas will be developed that we shall
use in subsequent discussion. When we have finished, you will not know how to
do a normal coordinate analysis, but hopefully you will have a rough idea of
what is involved.
Just as the electronic energy and electronic wave functions of a molecule are
related by a secular determinant and secular equations (Chapter 3), the vibrational
energies, vibrational wave functions, and force constants are related by a secular
determinant and a series of secular equations. The vibrational secular determinant
will be given here as (for the derivation see references 2 and 3):
F, 1 (G 1);, F 1 2 -(G- )l - F1 - (G- 1)I,

-0
1
F,1 (G )iA F. 2
- (G- )n 2
A ... F,, - (G -'),X
Here A = 4t 2 v2 , where v is the vibrational frequency (note the resemblance of 0

this GA term to the energy term in a molecular orbital calculation); this quantity
is known, for one begins by making a tentative assignment of all the normal
modes to the bands in the spectrum. The basis set for this calculation (i.e., the H H
counterpart of atomic orbitals in an m.o. calculation) is the set of internal Le
coordinates of the molecule expressed in terms of 3N - 6 atom displacements,
L. Three such coordinates, needed to describe the normal modes for the water FIGURE 6-7 Internal
coordinates for the H20
molecule, are illustrated in Fig. 6-7. The selection of these internal ccordinates
molecule.
is complicated in larger highly symmetric molecules because redundant coordinates
can result. One proceeds by selecting an internal displacement vector for a change
in bond length for each bond in the molecule and then selecting independent
bond angles to give the 3N - 6 basis set. The normal mode is going to be some
combination of this basis set of internal displacement coordinates, and the
vibrational wave function will tell us what this combination will be (again note
the resemblance of this to an electronic wave function for a molecule consisting
of the atomic orbital bases set). In the secular determinant given above, Fi1 is
the force constant for stretching the O-H bond along Lit, F2 2 is the force
constant for stretching along L 22 , and F, is related to the bend along L. The
off-diagonal element F2 1 in the vibrational secular determinant is called an
interactionforce constant, and it indicates how the two isolated stretches interact
with one another. When, for example, L 22 is subjected to a unit displacement,
the bond along L,1 will distort to minimize the potential energy of the strained
molecule. F2 1 is roughly proportional to the displacement of the bond along L
22
resulting from minimization of the energy of the molecule after displacement
along L,,. Part of the interaction relates to how the bond strength along L
22
changes as the oxygen rehybridizes when the bond along L, is stretched. Fij is
not necessarily identical to Fy.
The F-matrix elements account for the potential energies of the vibration.
The G-matrix elements contain information about the kinetic energy of the
molecule. The latter can be written exactly for a molecule from formulas given
by Wilson, Decius, and Cross(2 ) if we know the atom masses, the molecular bond
distances, and the bond angles.* This is a symmetric matrix. In practice, the force
constants are usually the only unknowns in the secular determinant, and they
can be determined. (Now the problem differs from the format of the molecular
orbital calculation.)
The secular determinant given above can be written in matrix notation as:
|F - G- 14l = 0
If we multiply by G we get
|GF - El = 0 (6-5)
* The evaluation of the G-matrix for a bent XY 2-molecule has been worked out in detail in
reference 18, Section 1-11. The reader is referred to this treatment for further details.
where E is the unit matrix. Furthermore, the G matrix has the property that the
's in the off-diagonal elements of the previous secular determinant have been
eliminated. The problem now is to solve
G11-- G1, F11 -- F , 1 --- 0 --- 0

1 =0
Gni Gnn Fn1 Fnn 0 1
where the G elements are all known, as are the 's. There are in general n values
of 2 that satisfy tiis-equation;-and theyare known. The difficulty with normal
coordinate analyses on most molecules is that thertare more unknown force
constants than there are frequencies. For the water molecule, the force constant
matrix is 3 x 3, and there are four unknowns* for the three frequencies. If one
deuterates the water molecule, new frequencies are obtained, but no new force
constants are introduced. The 160, ", and "O molecules could also be studied.
The problem here is whether or not the new equations we obtain in trying to
get more simultaneous equations than force constants are different enough to
allow for a unique, meaningful solution. (This is similar to the K - 1 vs. 8 problem
discussed in Chapter 4. Two parallel or nearly parallel equations do not solve
the problem, even though in the latter case a computer can pick a minimum.)
In the simpler molecules, this is often not a problem; but in more complex
molecules, the analysis is not to be believed unless errors are reported and some
statistical criterion of the significance of the fit is presented. This is readily
accomplished with a correlation matrix that indicates how extensively the
individual force constants are correlated. This problem is potentially so severe
that one should not accept the results from a force constant analysis unless the
correlation matrix is reported.
Suppose we were to do a normal coordinate analysis for Mn(CO),X. We
would have 30 internal displacement coordinates (3N - 6) and a 30 x 30 force
constant matrix, or many force constants. Symmetry can reduce this matrix to
smaller blocks by requiring certain interaction constants to be zero (vide infra),
but there remain several approximations that must be introduced to solve the
problem. In one analysis of the problem, one could assume that the carbonyl
stretches are so far removed in frequency from any of the other vibrations that
they can be treated separately.(') This is equivalent to claiming that the
metal-carbon stretch does not influence the C-0 vibration and setting all
interaction constants of the C--O stretches with anything else equal to zero. This
leads to a 5 x 5 block for the total secular determinant of force constants. Isotopic
substitution is then employed to solve the problem. In a more thorough analysis,( 5 )
the interaction constants between M-C and C-0 stretch coordinates that are
trans to each other have been determined and found to be significant.
Cotton and Kraihanzel(') have proposed a crude approximation for these
systems. They arbitrarily set any interaction force constant Frans for carbonyl
moieties that are trans to one another equal to 2Fai where F, corresponds to
interaction constants involving groups that are cis. The numbers of unknowns
and frequencies become comparable in many carbonyl compounds when this is
done. Jones(5 ) has carried out a more rigorous evaluation of this problem and
* The symmetry of the water molecule leads to this.

suggests that severe limitations be placed on systems for which the Cotton-
Kraihanzel approximations are invoked. It has been shown in the more complete
study that Fe, is considerably larger than Ftrans. Another approach that is being
investigated to facilitate normal coordinate analyses involves transferring inter-
action force constants from a simple molecule, in which they are well known, to
larger, similar systems.
Let us next consider a simple system, which we assume to have been solved
for meaningful force constants. The force constant values can be substituted into
the G and F matrices previously given, and the secular equations are now written
by multiplying by a matrix of the basis set L, the internal displacements; i.e.,
|GF -E E I|L I =0 (6-6)
In the case of water, F is 3 x 3, G is 3 x 3, E is 3 x 3 unit matrix, L is 3 x 1

(LI1 , L 2 2 , and L3 3), and ;, is the eigenvalue for which we wish to determine the
normal mode. The energies are now substituted one at a time. Matrix multiplications
then yield the contribution of the basis set (i.e., the contribution of the individual
internal displacements) to the vibrational wave function for the normal mode
corresponding to that frequency. This is done three times for the water molecule
with the three frequencies, leading to the wave functions for the three normal
modes.
For the water molecule, the wave functions for the 3N - 6 vibrations are:
1'I= NLO + N'(L 1 + L2 2 ) N N'

2 = N(L1 1 + L 2 2 ) + N'Le
1
/3 = (L - L )
_,2 11 22
The first two have A, symmetry, and the last has B, symmetry.
If we had a diatomic molecule (e.g., HC), all of the above matrices would
be 1 x 1 and we would have from equation (6-6)
FG -).= 0
where the G matrix element is the reduced mass and F 1 is the only unknown.
Rearranging, we see that
Fn 47r2V2
pHci
or
F1 I
v =(6-7)
The resemblance of this result to equation (6-3) is clear for AE = hv. Now we
see why equation (6-3) can be used to give the force constant for a bond in a
diatomic molecule but not in a more complex molecule. In larger, more complex
molecules, the observed frequency corresponds to a more complex vibration that

depends upon several bond force constants. The results of a normal coordinate
analysis produce the force constants and indicate exactly the form of each normal
mode in terms of the internal coordinates. For water, for example, the internal
coordinates in Fig. 6-7 are combined to produce normal modes that are similar
in form to those shown for SO2 in Fig. 6-5. This is expected because the internal
displacement coordinates (which are of the same form for SO2 and H 2 O) can
form the basis for producing a total representation by employing the symmetry
operations of the point group on them. Factoring the total representation
produces the symmetry of the irreducible representations, which irrthis-ease are
the normal modes. For the water molecule, the total representation b-taihed by
operating with the C,, point group on the internal displacements shown in Fig.
6-7 is 3 1 3 1 (the angle 0 is not changed by o or C 2 ). Factoring the total
representation produces 2A 1 + B 1 . Next, projection operators can be used on
this basis set just as on an a.o. basis set to produce
P(A1)LI ~ L,1 + L 2 2
P(A1 )LO Le (6-8)
P(B1)L 1 ~ L 1 - L2 2
This problem nicely illustrates the analogy between the symmetry aspects of the
molecular orbital problem and those of the vibrational problem.
The inverse of the force constant matrix produces a matrix of compliance
constants.(6) A diagonal compliance constant is a measure of the displacement
that will take place in a coordinate as a result of a force imposed upon this
coordinate, if the other coordinates are allowed to adjust to minimize the energy.
There are reported advantages to employing compliance constants instead of
force constants, and the reader is referred to the literature for details.(6 )
6-7 GROUP VIBRATIONS AND THE LIMITATIONS OF

THIS IDEA
The idea that we can look at the spectrum of a complex molecule and assign
bands in the spectrum to various functional groups in the molecule is called the
group vibration concept. This approach arose from the experimental observation
that many functional groups absorb in a narrow region of the spectrunm-regardless
of the molecule that contains the group. In the acetone molecule,* for example,
one of the normal modes of vibration consists of a C-O stretching motion with
negligible motion of the other atoms in the molecule. Similarly, the methyl groups
can be considered to undergo vibrations that are independent of the motions of
the carbonyl group. In various molecules it is found that carbonyl absorptions
due to the stretching vibration occur in roughly the same spectral region
(- 1700 cm '). As will be seen in the section on applications, the position of the
band does vary slightly ( 150 cm-1) because of the mass, inductive, and
conjugative effects of the groups attached. The methyl group has five characteristic
* This molecule has been analyzed by a normal coordinate treatment.

6-7 Group Vibrations and the Limitations of This Idea 161
absorption bands: two bands in the region from 3000 to 2860 (from the asymmetric
and symmetric stretch), one around 1470 to 1400 (the asymmetric bend), one
around 1380 to 1200 (the symmetric bend), and one in the region from 1200 to
800 cm-1 (the rocking mode, in which the whole methyl group twists about the
C-C bond, undergoing a rocking-chair type of motion). The concept of group
vibrations involves dissecting the molecule into groups and assigning one or
more bands in the spectrum to the vibrations of each group. Many functional
groups in unknown compounds have been identified by using this assumption.
Unfortunately, in many complicated molecules there are many group vibrations
that overlap, and assignment of the bands in a spectrum becomes difficult. It is
possible to perform experiments that help resolve this problem: deuteration will
cause a shift in vibrations involving hydrogen (e.g., C-H, O-H, or N-H
stretches and bends) by a factor of 1.3 to 1.4; characteristic shifts, which aid in
assignments, occur with certain donor functional groups (e.g., C=0) when the
spectrum is examined in hydrogen-bonding solvents or in the presence of acidic
solutes.
The limitations of the group vibration concept should be emphasized so that
incorrect interpretations of infrared spectral data can be recognized. Thegroup
vibration concept implies that the vibrations of a particular functional group are
re~Titvely Iidependent of the rest of the molecule. If the center of mass is notto
move, this is impossible. All of the nuclei in a molecule must undergo their
harmonic oscillations in a synchronous manner in normal vibrations. In view of
the discussion fi-fhe l5revidus sc'tion, we can see that the group vibration concept
will be good when the normal mode wave function consists largely (80 to 90%)
of the internal displacement in which only the group is involved, i.e., the functional
group vibration is the main internal displacement coordinate. This will occur
when the symmetry properties of the molecule are such as to restrict the
combination of the internal displacements of the group with otherirternal
displacement coordinates; e.g., L, in the water molecule cannot contribute to the
B1 asymmetric stretch. When the vibrational motions involving the two internal
displacement coordinates are very different in energy, then the off-diagonal
interaction constants are small, the two internal displacement coordinates are
not mixed extensively in the vibrational wave functions, and the vibrations can
be treated separately as group vibrations. This is equivalent to the assumption
made when we treated the carbonyl vibrations in M(CO),X separately to obtain
the 5 x 5 matrix.
When the atoms in a molecule are of similar mass and are connected by
bonds of comparable strength (e.g., all single bonds as in BF 3 -NH 3 ), all the
normal modes will be mixtures of several internal displacement vectors. For
example, in BF 3 -NH 3 it would be impossible to assign a band to the B-N
stretching vibration because none of the normal modes correspond predominantly
to this sort of motion. When this occurs, the various group vibrations are said
to be coupled. This term simply describes a situation in which the very crude
group.vibration concept is not applicable to the description of the normal mide
corresponding to a particular absorption band. This discussion should be reread
if the reader cannot distinguish coupling from the phenomena of Fermi resonance
or combination bands.
The HCN molecule is interesting to consider in the context of the above
discussion. The frequencies of a pure C-H and pure C-N stretch are similar
in energy. The C-H and C-N group stretching vibrations in the H-C-N
molecule are coupled. The absorption band attributed to the C-H stretch actually
involves C-N motion to some extent and vice versa; i.e., the observed frequencies
cannot be described as being a pure C-H and a pure C-N stretch. Evidence
to support this comes from deuteration studies, in which it is found that
deuteration affects the frequency of the band that might otherwise be assigned
to the C-N stretch. This absorption occurs at 2089 and 1906 cm-' in the
molecules H-C-N and D-C-N, respectively. Equation (6-3) would predict
that deuteration should have very little effect on the C-N frequency, for it has
little effect on the C-N reduced mass. The only way the frequency can be affected
is to have a coupling of the C-D stretch with the C-N stretch. Upon deuteration,
the C-H stretch at 3312 cm-1 is replaced by a C-D stretch at 2629 cm-1.
Since the C-D frequency is closer to the C-N frequency than is that of C-H,
there is more extensive coupling in the deuterated compound. Because the
symmetries of the normal modes differ, the C-D bending vibration does not
couple with the "C-N stretch."
It would be improper to draw conclusions concerning the strength of the
C-H bond from comparison of the frequency for the C-H stretch in H-CN
with data for other C-H vibrations. The force constants for the C-H bonds
can be calculated for the various compounds by a normal coordinate analysis.
These values should be employed for such comparisons.
As another example of the difficulties encountered in the interpretation of
frequencies, consider the stretching vibration for some carbonyl groups. Absorp-
tion bands occur at 1928, 1827, and 1828 cm-1 in F 2 CO, CI2 CO, and Br 2 CO,
respectively.(7 ) One might be tempted to conclude from the frequencies that the
high electronegativity of fluorine causes a pronounced increase in the C-O force
constant. However, a normal coordinate analysis(7 ) shows that the C-O and
C-F vibrations are coupled and that the C-O stretch also involves considerable
C-F motion. The frequency of this normal mode is higher than would be expected
for an isolated C-O stretching vibration with an equivalent force constant. There
is a corresponding lowering of the C-F stretching frequency. Since chlorine and
bromine are heavier, they make little contribution to the carbonyl stretch. The
normal coordinate analysis indicates that the C-O stretching force constants
are 12.85, 12.61, and 12.83 millidynes/A in F 2 CO, Cl2 CO, and Br 2CO, respectively.
A normal coordinate analysis and infrared studies of 1 2 C-, 3 C-, 160-, 180-,
and H- and D-substituted ketones containing an ethyl group attached to the
carbonyl group indicate(8 ) that the band near 1750 cm-1 assigned to the C-O
stretch in these ketones corresponds to a normal mode that consists of about
75% C-O and 25 % C-C stretching motions. These examples should be kept
in mind whenever one is tempted to interpret the frequencies of infrared bands
in terms of electronic effects. Force constants obtained from normal coordinate
analyses should be compared; frequencies read directly from spectra should not
be. Unfortunately, reliable force constants are in limited supply.
Raman 6-8 INTRODUCTION

Spectroscopy Raman spectroscopy is concerned with vibrational and rotation'al transitions,
and in this respect it is similar to infrared spectroscopy. Since the selection rules
are different, the information obtained from the Raman spectrum often comple-
Laser
source
FIGURE 6-8 The Raman

experiment (schematic).
Detector
ments that obtained from an infrared study and provides valuable structural
information.
In a Raman experiment, a monochromatic beam of light illuminates the
sample, and observations arqemadgon the scattered lght .rjghtangs oqhe
incj6deaLhagw .8k), Monochromatic light sources of different frequencies
can be employed; gas lasers are commonly used for this purpose. Various sample
holder configurations are employed"'; the only requirement is that the detector
must be at right angles to the source. This configuration is illustrated in Fig. 6-8
fo-ra ca~pillary cell. Ab(orition of the monochromatic light beam, leading to
decomposition, can be a problem, as can fluorescence. These problems are
minimized by choice of an appropriate gas laser"")line: the He-Ne laser gives
a line at 6328 A (red); the Ar laser gives lines at 4579, 4658, 4765, 4880, 4915,
and 5145 A (blue-green) (the Art 4880 A and 5145 A lines are usually used); and
the Kr laser gives 5682 and 6471 A lines. The introduction of tunable dye lasers
has extended the range of usable wavelengths even further. Use of a narrow
bandpass filter for each laser line reduces effects from the large number of laser
ghosts. A valuable discussion of various aspects of this experiment is contained
in reference 11.
Center of a
broad
Rayleigh
line
FIGURE 6-9 Lines observed in a

Stokes Anti-Stokes Raman spectrum (schematic).
line line
0
nc- Veas V-+V
Increasing V
We should wonder how we obtain vibrational transitions when we employ

0
a visible or u.v. source. A quantum of the incident light of frequency v and
0
energy hv can collide with a moleculland be scattered with unchanged frequency.
This is referred to as Rayleigh scattering. The mechanism involves inducing a
dipole moment, D, in the molecule when it is in the field of the electric vector of
the radiation. The electrons in the molecule are forced into oscillations of the
same frequency as the radiation. This oscillating dipole radiates energy in all
directions and accounts for the Rayleigh scattering. If the photon is actually
absorbed in the process and re-emitted, this is fluorescence. The difference between
scattering and fluorescence is thus a subtle one, having to do with the lifetime
of the species formed in the photon-molecule collision.
The scattering process just described corresponds to an elastic collision of
the molecule with the photon. Ranan-scattering involves an inelastic collision.
The molecule in its ground vibrational state accpts energy from theopbfoon
being scattered, exciting the molecule into a higher vibrational state, while the
incident radiation now becomes scattered with energy h(v - v.). In the Raman
spectrum of the scattered light measured at right angles, we now detect a frequency
V - v., called the Stokes line, as shown in Fig. 6-9. The measured value of v, is
identical to the infrared frequency that would excite this vibrational mode if it
were infrared-active.
A molecule in the vibrationally excited state v = I can collide with an incident
light quantum of frequency v0 . The molecule can return to the ground state by
giving its additional energy hv, to the photon. This photon, when scattered, will
have a frequency v + v,. The spectral line with this frequency in the Raman
spectrum is referred to as an anti-Stokes line (see Fig. 6-9). Because of the
Boltzmann distribution, there are fewer molecules in the v = 1 state than in v = 0,
and the intensity of the anti-Stokes line is much lower than that of the Stokes line.
Both Rayleigh and Raman scattering are relatively inefficient processes. Only
about 10-3 of the intensity of the incident exciting frequency will appear as
Rayleigh scattering, and only about 10-6 as Raman scattering. As a result, very
intense sources are required in this experiment. Laser beams provide the required
intensity (- 100 milliwatts to ~1 watt of power) and produce good spectra even
with very small samples.
6-9 RAMAN SELECTION RULES
The molecule interacts with electromagnetic radiation in the Raman experiment

via the oscillating induced dipole moment or, more accurately, the oscillating
molecular polarizability. The intensity, I, of the scattered radiation, v, is a function
of the polarizability, aj, of the molecule which, for a randomly oriented solid, is
given by:
32 c4 I5 v 4s a (6-9)
where I is the intensity of the incident radiation and ac is an element of the

molecular polarizability tensor. Molecular polarizability can be represented by
an ellipsoid; the change in polarizability (as the CO 2 molecule, for example,
undergoes the symmetric stretch) can be schematically illustrated as in Fig. 6-10.
6- Raman Selection Rules 165
symmetric
stretching -O m Cil1llll O-z -O i C O-z -0 C O11
0-z
displacements (extended)
(equilibrium) (compressed)
corresponding
polarizability
ellipsoid
-z - z
FIGURE 6-10 The change in polarizability for the C02 molecule for various
displacements of the atoms during the symmetric stretching vibration. The
double-headed arrows in the ellipsoid indicate (a) the z-axes of the ellipsoid, (b)the
x, (c) the y, and (d)the xz.
When a molecule is placed in a static electric field, the nuclei are attracted
toward the negative pole and the electrons toward the positive pole, inducing a
dipole moment in the molecule. If D represents the induced dipole moment and
E the electric field, the polarizability a is defined by the following equation:
D = aE (6-10)
The magnitude of this polarizability depends upon the orientation of the bonds
in the mokeukwith respect to the electric field direction; i.e., it is anisotropic.
In most molecules it is a tensor quantity. The induced dipole moment in the
molecule will depend upon the orientation of the molecule with respect to the
field and, since the dipole moment is a vector, it will have three components D.,
D,, and D.. The polarizability tensor can be represented physically as an ellipsoid
(see Fig. 6-10), and it has nine components that can be described by a 3 x 3
matrix. The induced dipole moments are related to the polarizability tensor by
the following equations:
Dx = xxE + cxE, + xzEz
D, = x,,XE, + a,, E, + x,, E. (6-11)

D, = xzxEx + xt, E, + xz E,
These equations state that if we locate a molecule in an electric field so that the
ellipsoid axes point between the Cartesian axes, then the x-component of the
induced dipole moment will depend upon the magnitude of the electric field along
the y-axis because there is a polarizability component along xy that can interact
with E, and make a contribution to the change observed along x. The same is
true for the interaction of E. with 2,. Equations (6-11) can be written in matrix
form as
Dj = a,,X O,, OXz_ Ej

DY azX a , aCz Ez
Foran optically inactive molecule, the a tensor is symmetric and
If we choose a proper set of axes, we can carry out a transformation that

diagonalizes the tensor previously given. The new axes are called the principal
axes of the polarizability tensor with components 2,x, Oy, and 2-,, whose trace
is equal to that in the previous coordinate system.
Quantum mechanically, the polarizability of a molecule by electromagnetic
radiation along the direction ij is given by
- L(IM)),
h , v, -v
(MV)me(Mj)e]
v, + v,
(6-12)
where m and n are the initial and final states of the molecule and e represents
an excited state. Mi and Mi are electric dipole transition moments along i and
j, whereas v, is the energy of the transition to e and v, and v, are the frequencies
of the incident and scattered radiation.
As previously mentioned, the selection rules for Raman spectroscopy are
different from those for infrared. In order for a vibration to be Raman-active, the
change in polarizabilityof the molecule with respect to vibrationalmotion must not
be zero at the equilibrium position of the normal vibration; i.e.,
- #0r, 0 (6-13)
where a is the polarizability and r represents the distance along the normal
coordinate. If the plot of polarizability, 2c,versus distance from the equilibrium
distance, re, along the normal coordinate is that represented by Fig. 6-11(A), the
vibration will be Raman active. If the plot is represented by the curves 1 or 2 of
Fig. 6-11(B), ac2/ar will be zero at or near the equilibrium distance, re, and the
vibration will be Raman-inactive.
Small amplitudes of vibration (as are normally encountered in a vibration
mode) are indicated by the region on the distance axis on each side of zero and
between the dotted lines. As can be seen from Fig. 6-11, the vibration in (A)
6-9 Raman Selection Rules 167
/C11
I
re+ - re +
distance r distance r
(A) (B)
FIGURE 6-11 Polarizability as a function of distance for some
hypothetical molecules.
corresponds to an appreciable change in polarizability occurring in this region,

while that in (B) corresponds to practically no change. Hence, the Raman selection
rule (6-13) is often stated: In orderfor a vibration to be Raman-active, there musts
be a change in polarizability during the vibration.
The completely symmetric stretch for CO 2 is represented in Fig. 6-10. The
points labeled A and C in Fig. 6-11(A) correspond to extreme stretching of the
bonds for the symmetric stretch. A and C represent, respectively, more ard less
polarizable structures than the equilibrium structure. A plot of a versus r for the
symmetric stretch is represented by Fig. 6-11(A). The change in polarizability for
the asymmetric stretch is represented by one of the curves in Fig. 6-11(B). As a
result, this vibration is Raman-inactive. Infrared activity is just the opposite of
Raman activity in this case. In general,for any molecule that possesses a center) T
of symmetry, there will be no fundamental lines in common in the infrared and /
Raman spectra. This is a very valuable generalization for structure determination.
If the same absorption band is found in both the infrared and Raman spectra,
it is reasonably certain that the molecule lacks a center of symmetry. It is possible
that in a molecule lacking a center of symmetry no identical lines appear because
of the low intensity of one of the corresponding lines in one of the spectra.
This discussion serves to illustrate qualitatively the selection rules for Raman
spectroscopy. It is often difficult to tell by inspection the form of the polarizability
curve for a given vibration. As a result, this concept is more difficult to utilize
qualitatively than is the dipole moment selection rule in infrared. In a later section
it will be shown how character tables and symmetry arguments can provide
information concerning the infrared and Raman activity of vibrations. For now,
we summarize the above discussion by stating that in order for a vibration to
be Raman active, the shape, size, or orientation of the ellipsoid must change
during the vibration, as shown in Fig. 6-11(A).
Oscillating induced
dipole
Spherical
isotropic
7
FIGURE 6-12 Raman scattering of yz plane-polarized radiation producing an
xz-polarized scattered wave when the polarizability is isotropic.
6-10 POLARIZED AND DEPOLARIZED RAMAN LINES
Valuable information can be obtained by studying the polarized components of

a Raman line. In order to understand how the Stokes lines can have a polarization,
we will first consider some scattering experiments with polarized light. In Fig.
6-12 the interaction of yz-polarized radiation with a molecule that has an isotropic
polarizability tensor is indicated. The oscillation induced in the molecule will be
in the same plane as the electric field and will have a z-component as shown by
the double-headed arrow. If we observe emitted radiation along the x-axis, it will
have to be xz polarized. Since ot is isotropic, the direction of the induced dipole
is unchanged as the molecule tumbles, and only xz scattered radiation results. If
we had used xy-polarized radiation as our source, the oscillating induced dipoles
in a molecule with an isotropic polarizability would be oscillating parallel to the
x-axis. In Fig. 6-13 the x-axis is perpendicular to the page, so the xy-plane
polarized radiation is also perpendicular to the page. No component is detected
z Oscillating induced dipole
YX
FIGURE 6-13 Raman scattering of xy-plane polarized radiation
produces no xz or yz components for the scattered wave when the
polarizability is isotropic.
6-10 Polarized and Depolarized Raman Lines 169
along x, for the molecular induced dipole moment oscillates parallel to x for all
orientations of the molecule.
As a result of the above discussion, it should now be clear that-even if our
source is non-polarized-the scattered radiation will be polarized if our molecule
is isotropic in a, for there will be no xy-component.
The polarizability tensor is usually anisotropic, as shown in Fig. 6-14.
Accordingly, the induced moment will not be coincident with the plane of the
electric field but will tend to be oriented in the direction of greatest polarizability.
The scattered light vibrates in the same plane as the induced dipole. As the
molecule tumbles, the orientation of the induced dipole moment relative to the
x-axis changes. Therefore, there are both xz and xy-components in the scattered
beam giving rise to radiation that is depolarized. Even if the incident radiation
is polarized, an anisotropic polarizability tensor will give rise to scattered radiation
that is depolarized. In the actual experiment, polarized radiation from a laser
source is employed and the polarization of the radiation scattered along the
x-axis is determined.
Non-polarized radiation
FIGURE 6-14 Raman scattering of non-polarized radiation from an

anisotropic polarizability tensor.
The above considerations apply to the Stokes lines, for which it is found
that a totally symmetric vibration mode gives rise to a polarized scattered line and
that a vibration with lower symmetry is depolarized. This property can be used to
confirm whether or not a vibration has A, symmetry. Thus, by using an analyzer,
the scattered Stokes radiation traveling along the x-axis is resolved into two
polarized components: radiation polarized in the y-direction and that in the
z-direction. Radiation traveling along x and polarized in the y-direction is referred
to as perpendicular and that component polarized in the z-direction is called
parallel. The depolarizationratio, p, is defined as the ratio of the intensity, I, of
the perpendicular (y(L)) to the parallel, z(I), components of the Stokes line:
p = 'Y-) (6-14)
Those Raman lines for which p = 3/4 are referred to as depolarized lines and
correspond to vibrations of the molecule that are not totally symmetric. Those
Raman lines for which 0 < p < 3/4 are referred to as polarized lines. The vibrations
of the Mkzuk~aust ormas An orderoq e lrized The use of this
information in making band assignments will be made clearer after a discussion
of symmetry considerations.
6-11 RESONANCE RAMAN SPECTROSCOPY
As the laser exciting frequency in the Raman experiment approaches an allowed

electronic transition in the molecule being investigated, those normal modes that
are vibronically active in the electronic transition exhibit a pronounced enhance-
ment in their Raman intensities. ") There are two applications of this fact. One \
involves utilizing the intensity enhancement to study materials that can be obtained
only at low concentrations. The second application involves the fact that only
those few vibrational modes near the site of an electronic transition in a complex
molecule will be enhanced. Most examples of resonance Raman spectroscopy
involve the enhancement of totally symmetric modes and arise from intense
electronic transitions. There are recent examples in which vibronically allowed
electronic transitions enhance the Raman bands of modes that are not totally
symmetric, making them vibronically allowed.
It is possible for more than one electronic transition to be responsible for the
resonance Raman effect, complicating the symmetry-based application just
discussed. By studying the intensity of the resonance Raman bands as a function
of the frequency of the exciting laser line, one can determine whether one electronic
transition (the so-called A-mechanism) or two (B-mechanism) are involved. The
intensity of a particular band is proportional to the square of the frequency factor,
F. For the two mechanisms above, equations (6-15) and (6-16) have been proposed,
respectively:
FA (V,2 + V0 2 ) (6-15)
A y2
V _ 02 22
where v, is the frequency for the electronic transition, v, is the source frequency
and v = v, - Av.n with Av,. equal to the frequency for the vibrational transition;
and
F = (Ve
2 2
)(v,
+ v2 ) (6-16)
* v2_ Vo2)(V,2 _ VT
2
where ve and v, are the frequencies for the two electronic transitions involved and
the other terms are as defined for equation (6-15). A large number of other
mechanisms and equations have recently been proposed.
An example of the enhanced intensityt(3) from the resonance Raman effect is
shown in Fig. 6-15. In (A), a typical electronic spectrum for a heme chromophore
is shown. In (B), parallel and perpendicular components of the resonance Raman
spectrum of a 5 x 10-4 M solution of oxyhemoglobin are illustrated. Without the
resonance enhancement associated with the 568 nm exciting wavelength, this
concentration of material would not have yielded a Raman spectrum. The label
dp refers to a depolarized band (depolarization ratio of 3/4), p to a polarized band,
and ip to a band in which the depolarization ratio is greater than one (i.e., the
scattered light is polarized perpendicular to the polarization of the incident light).
An ip band requires that the scattering tensor be antisymmetric (i.e., aj = -1ji),
which is possible when vo approaches ve. When the exciting line is far from an
electronic transition, the ratio cannot be larger than 3/4. The scattering tensor
elements discussed here and symbolized by a are not to be confused with the
polarizability tensor elements discussed earlier.
An assignment of the heme Raman spectrum has been made that is consistent
with the assigned electronic transitions. The spectral features common to many
6-11 Resonance Raman Spectroscopy 171
30 -XVisible (a) resonance

(A) (A) 30- ~ X5 i
2-Soret--|
20- resonance
mM
10-
0
350 400 450 500 550 600
X(my)
OD W) Hb02
0
~o o
I - 9~ 'N"t*8 ~ d pII
p 'Po p2 01
dp dp
in I iI i I d I I I I [
600 500 1400 300 1200 1100 1000 900 aoo 700 600
(B) CM1
FIGURE 6-15 (A) Ultraviolet (Soret) and visible (c -#)
chromophores in hemes. Spectrum of ferrocytochrome c. (B)
Resonance Raman spectra of oxyhemoglobin. Both the direction
and the polarization vector of the incident laser radiation are
perpendicular to the scattering direction. The scattered radiation
is analyzed into components perpendicular (/i) and parallel (i)
to the incident polarization vector. The exciting wavelength was
568.2 nm (5682 A) for HbO 2. Concentrations are about 0.5 mM for
HbO 2 . [From T. G. Spiro and T. C. Strekas, Proc. Natl. Acad. Sci.,
69, 2622 (1972).]
heme proteins are discussed, as is the use of the depolarization ratios to infer local
symmetry of the heme unit.! 3 )
In another application of this technique, 14) a resonance Raman analysis of
oxyhemerythrin was carried out using the oxygen -+ iron charge-transfer band at
5000 A to intensify the 0-0 stretching vibrations. Two Raman frequencies were
observed at 844cm (p = 0.33) and at 500cm- '(p = 0.4). When "02 is employed,
the bands shift to 798 cm and 478 cm , respectively. These bands do not appear
in the deoxygenated protein. The band at 844 cm-' has been assigned to the
0-0 stretch and is not coupled to the other vibrational modes in the molecule,
since the predicted frequency for isotopic substitution (796 cm') is observed. The
500 cm- 'band is assigned to an Fe-O-0 stretching mode. The spectra indicate
only one type of 02 complex.
Symmetry 6-12 SIGNIFICANCE OF THE NOMENCLATURE USED TO

DESCRIBE VARIOUS VIBRATIONS
Aspects of
Molecular In this section the conventions used to label the vibrations A 1, A 2 , etc., will be
Vibrations reviewed to refresh the reader's memory. If the vibration is symmetric with respect
to the highest-fold rotation axis, the vibration is designated by the letter A. If it is
antisymmetric, the letter B is employed. E stands for a doubly degenerate vibration
and T for a triply degenerate mode (F is commonly used instead of T in vibrational
spectroscopy). The subscripts g and u refer to symmetry with respect to an inversion
through a center of symmetry and are used only for molecules with a center of
symmetry. If a vibration is symmetric with respect to the horizontal plane of
symmetry, this is designated by ', while " indicates antisymmetry with respect to
this plane. Subscripts 1 and 2 (as in A, and A 2 ) indicate symmetry and antisymmetry,
respectively, toward a twofold axis that is perpendicular to the principal axis.
6-13 USE OF SYMMETRY CONSIDERATIONS TO

DETERMINE THE NUMBER OF ACTIVE INFRARED
AND RAMAN LINES
In this section we shall be concerned with classifying the 3N - 6 (or 3N - 5 for

a linear molecule) vibrations in a molecule to the various irreducible representations
in the point group of the molecule. This information will then be used to indicate
the degeneracy and number of infrared and Raman-active vibrations. The procedure
is best illustrated by considering the possible structures for SF 4 (C2 c, C3 ,, and T)
represented in Fig. 6-16.
We first consider the C2 structure. An x, y, z-coordinate system is constructed
for each atom, and we proceed to determine the total representation for this basis
set using the C2 character table reproduced in Table 6-4.
We know that when an atom is moved by a symmetry operation, the
contribution to the total representation will be zero; so we need only worry about
the x, y, and z coordinates on atoms not moved. For E, none of the atoms or
coordinates are moved, giving XT(E) = 15. Only the sulfur atom is not moved by
the C 2 operation, but the x and y coordinates of the sulfur each contribute -1
to the total representation while the z contributes + 1. This leads to Xr(C 2 ) = -1.
Sulfur and two of the fluorines are not moved by reflection in the or, plane, while
two fluorines are moved. On the sulfur atom this reflection changes the sign of y
but leaves x and z unchanged for a contribution of + 1. The same is true for each
F TABLE 6-4. C2V Character Tablea

.4
'S- E C2 0m z
F-j '-F Al 1 1 1 1 z , 2
F A 2 1 1 -l 1 Rz, 2[,
B1 1 1 1 1 x, R,, 2,
Td
B2 1 1 -1 1 yR
FIGURE 6-16 Some possible
"For planar molecules the x-axis is perpendicular to the plane.
structures for SF4.
6-13 Use of Symmetry Considerations to Determine the Number of Active Infrared and Raman Lines 173
of the unshifted fluorines, leading to YT(xZ)) = 3. Reflection through yz changes

the x-coordinate to - 1 for all three atoms not moved but does not change y or
z coordinates, leading to Zr(u',z)) 3. Summing up these results, we obtain:
T 15 -1 3 3
We can factor the total representation as follows:
for E for C2 for u(XZ) for a
nA1 [-1-15+1I--1(-) + 1-1.3 +1-1-3) =5

4
A1
nA= 4 [1-1-15+ 1-1-(-1) + 1-(-1)-3+ (-1)-3] 2
nBI=-[1-1-15+1 -(-1)( 1)+l-(-)-3 + 11-3] 4

4
n2 = -[I- - 15 + 1I(-)( 1 + 1 - 1.-3 + 1 -(--1) -3] =4

4
The results nA = 5, n -2 = 2, nB = 4, and n2 = 4 repesent the 3N degrees of freedom

for a pentatomic C2, structure.
To get the total number of vibrations, the three degrees of translational freedom
and three of rotational freedom must be subtracted (giving 3N - 6 for a nonlinear
molecule). Translation along the x-axis can be represented by an arrow lying along
this axis, which, as seen from the C 2 character table, belongs to the species B1.
Simiarly, translation along the z and y axes transforms according to species A,
and B 2 . Rotations along the x, y, and z axes are represented by the symbols Rx,
R, and R2 and belong to species A 2 , B, and B2 , respectively. TJhese six degrees
of freedom (1 A, 2B, and 2B 2 ) are subtracted from the representation of the
total degrees of fretedom, producing nA = 4 nA = 1, n"- 2, and nB = 2, corre-
sponding to the nine vibrations predicted from the 3N - 6 rule.
We have now calculated the species to which the 3N - 6 vibrations of SF4
belong if it has a C 2 , structure. Four are of species A 1, one of A2, two of B1, and
two of B 2 . The next job is to determine which vibrations are infrared active and
which are Raman active. For a fundamental transition to occur by absorption of
infrared electromagnetic radiation, one of the three integrals
f x& * dr f 5,ofrex d , *exdr (6-17)
must be non-zero or A 1 The c,9, and z operators correspond to the orientation

of the electric field vector relative to a molecular Cartesian coordinate system. The
operator is identical to that discussed in electronic absorption spectroscopy. Since
all ground vibrational wave functions are A 1 (if necessary, reread Chapter 4, the
section on atomic and molecular transitions), this amounts to having a component
of the transition dipole operator i, , or 2 that has the same symmetry as ,rib"e,
where the total vibrational excited state for a fundamental transition has the same
symmetry as the normal mode excited. Accordingly, afundamental will be infrared
active if the excited normal mode belongs to one of the irreducible representations
corresponding to the x, y, and z vectors, and will be inactive if it does not.
In order for a vibration to be Raman active, it is necessary that one of the
integrals of the type
f drexr (6-18)
be non-zero, i.e., the integrand A. Here the operator P is one of the quadratic
functions of the x, y, and z-vectors (i.e., x 2 , y2 , z2, xy, yz, xz), simply or in combination
(i.e., x 2 _ y 2 ). The symmetries of these functions are listed in the character tables
opposite their irreducible representations. Since these quantities are components
of the polarizability tensor, we find the following rule stated: Afundamental transition
will be Raman active if the correspondingnormal mode belongs to the same irreducible
representationsas one or more of the components of the polarizability tensor.
For the C 2 structure of SF 4 , the A 1, B 2 , and B, modes are infrared active,
while A 2 is infrared inactive (neither x, y, nor z has A2 symmetry). Eight infrared
bands are expected: 4Ai, 2B 1, and 2B 2 . All nine of the fundamental vibrations are
Raman active, and the four A, modes will be polarized.
When we carry out the above procedure on-the Td structure for SF 4 , we obtain
a total representation Z7 = 15(E) 0(8C 3 ) 3 (6 0-d) - 1(6S 4 ) - 1(3S 4 2 = 3C 2). Factor-
ing, we obtain:
n^,= [1-1-15+8-1-0 +6-1-3 +6-1-(-1) +31- (-1)] =1

24
nA2 = -[1-1-15+8-1-0 +6-(- 1)-3+6(- 1)(- 1)+3-1--1)] =0

24
1
nE= [1-2-15 + 8-(-1)0 + 60-3 +6-0-(-1) + 3-2(-1)] =1
24
1
nT=- [1-3-15 + 8-00 +6-(-1)3 +6-1-(-1) + 3-(1)-(1)]=1
24
1
N - [1-3-15 + 8-00 +6-13 +6- (-1)(-1) + 3-(-1)-(- 1)] 3
24
The T character table indicates that the three translations are of species T2
and the three rotations of species Ti. Since T, and T2 are triply degenerate, we
need subtract only one of each from the total degrees of freedom to remove three
degrees of translation and three of rotation. The total number of vibrations belong
to species A1, E, and 2T 2 giving rise to the nine modes predicted from the 3N - 6
rule. The six T2 vibrations (two triply degenerate sets) are infrared active and give
rise to two fundamental bands in the infrared spectrum. All modes are Raman
active, giving rise to four spectral bands corresponding to fundamentals. Of these,
A, is polarized. It can be shown similarly that the C,, structure of SF4 leads to
six infrared lines (three A1 and three L) and six Raman lines (three A 1 and three
E), three of which are polarized (A1).
6-13 Use of Symmetry Considerations to Determine the Number of Active Infrared and Raman Lines 175
TABLE 6-5. Summary of Active Modes Expected for Various Configurations
of SF4
C2 T, C, Found"
Infrared-active
modes 8(4A 1, 2B 1, 2B 2 ) 2(2 T2) 6(3A 1, 3E) 5(or 7)
Raman modes 9(4A 1, A 2 , 2B 1, 2B 2 ) 4(A 1 , E, 2T 2 ) 6(3A 1, 3E) 5(or 8)
Polarized modes 4(4A 1) 1(A1 ) 3(3A 1) I
"R. E.Dodd, L. A.Woodward, and H. L. Roberts, Trans. Faraday Soc., 52, 1052 (1956).
The actual spectrum is found to contain five infrared fundamentals, and five
Raman lines, one of which is polarized. These results, summarized in Table 6-5,
eliminate the Td structure, leaving either C,, or C, of the possible structures
considered. This example demonstrates the point that although there cannot be
more fundamental vibrations than allowed by a symmetry type, often not all
vibrations are detected. The problems in application of these concepts involve
separating overtones and combinations from the fundamentals and, as is often the
case, not finding some "active" fundamental vibrations because they have low
intensities.
Table 6-5 should be verified by the reader by using the character tables
contained in Appendix A and the procedure described above. By a detailed analysis
of the band contours and an assignment of the frequencies, it was concluded that
the C2, structure best fits the observed spectrum. This structure was subsequently
confirmed by other physical methods.
The data available on CHCl, and CDCl3 are summarized in Table 6-6 and
will be discussed briefly because this example serves to review many of the
principles discussed. The 3N - 6 rule predicts nine normal vibrations. Just as
with the C3, structure of SF 4, the total representation shows that these nine
vibrations consist of three A, and three E species or a total of six fundamental
frequencies. All six fundamentals will be IR and Raman active, and the three A,
fundamentals will give polarized Raman lines. The observed bands that have
been assigned to fundamental frequencies are summarized in Table 6-6. These
data indicate that the 3033, 667, and 364 cm-' bands belong to the totally
symmetric vibrations (species A,), for these bands are Raman polarized. They
are labeled v,, v2, and v3 , respectively. The other three bands are of species E,
and the 1205, 760, and 260 cm-' bands are v4 , v5 , and v6, respectively. Note
that since the molecule does not have a center of symmetry, the IR and Raman
TABLE 6-6. Infrared and Raman Fundamentals for CHCl 3

IR-Active Raman-Active
Vibrations for Vibrations for Raman Spectra
CHCl 3 (cm ') CHC13 (cm- ') of CDC13 (cm- ) Designation
260 262 262 v'

364 366 (polarized) 367 v3
667 668 (polarized) 651 v2
760 761 738 15
1205 1216 908 v4
3033 3019(polarized) 2256 VI
spectra have lines in common. The exact form of these fundamental vibrations
is illustrated in Appendix C. The bands vi and v 4 are the C-H stretching and
bending vibrations, respectively. Note how deuteration has a pronounced effect
on these frequencies but almost no effect on the others. The vibrations at 3033
and 1205 cm- are almost pure C-H modes. Since the very light hydrogen or
deuterium atom is moving, very little C-Cl motion is necessary to retain the
center of mass (to prevent net translation of the molecule).
The above manipulations are quite simple and yet yield valuable information
about the structures of simple molecules. There are many applications of these
concepts. It is strongly recommended that the reader carry out the exercises at
the end of this chapter.
Spectroscopists often use the v, symbolism in assigning vibrations. This can
be translated into the language of stretches, bends, and twists by referring to the
diagrams in Appendix C or in Herzberg (see Additional References) for an example
of a molecule with similar symmetry and the same number of atoms. The
irreducible representation of a given normal vibration can be determined from
the diagrams by the procedure outlined in Chapter 2.
6-14 SYMMETRY REQUIREMENTS FOR COUPLING

COMBINATION BANDS, AND FERMI RESONANCE
There are some important symmetry requirements regarding selection rules for
overtone and combination bands. These can be demonstrated by considering
BF 3, a D3,, molecule, as an example. The D3 h character table indicates that the
symmetric stretch, vi, of species A 1 ' is infrared inactive (there is no dipole moment
change). The species of the combination band v1 + v 3 (where v3 is of species E')
is given by the product of A1 ' x E' = E'. The combination band is infrared active.
The v2 vibration is of symmetry A 2" and is infrared active. The overtone 2v 2 is
of species A 2 " x A 2 " = A 1 ', which is infrared inactive; 3v2 is of species A 2 " and
is observed in the infrared spectrum. This behavior is strong support for the
initial structural assignment, that of a planar molecule, and serves here as a nice
example to demonstrate the symmetry requirements for overtones and combi-
nation bands (recall the discussion in Section 6-5).
In the case of Fermi resonance it is necessary that the states which are
participating have the same symmetry. For example, if 2v 2 is to undergo Fermi
resonance with vj, one of the irreducible representations for the direct product
2v 2 must be the same as for vi.
Coupling of group vibrations was mentioned earlier in this chapter. In order
for coupling to occur, the vibrations must be of the same symmetry type. For
example, in acetylene the symmetric C-H stretching vibration and the C-C
stretching vibrations are of the same symmetry type and are highly coupled. The
observed decrease in the frequency of the symmetric C-H stretch upon
deuteration is smaller than expected, because of this coupling. Since the
asymmetric C-H stretch and the "C-C"stretch are of different symmetry, they
do not couple, and deuteration has the normal effect on the asymmetric C-D
stretch (a decrease from 3287 to 2427 cm -1 is observed).
6-15 Microwave Spectroscopy 177
6-15 MICROWAVE SPECTROSCOPY
Pure rotational transitions in a molecule can be induced by radiation in the far

infrared and microwave regions of the spectrum. Extremely good precision for
frequency determination is possible in the microwave region. Compared to the
infrared region of the spectrum, where measurements are routinely made to about
1 cm- , one can get resolution of about 10 8 cm~1 in the mirowave region. A
wide spectral range plus resolution and accuracy to 10 - cm- ' make this a very
valuable region for fingerprint applications. A list of frequencies has been
tabulated" 5' 6 ) consisting of 1800 lines for about 90 different substances covering
a span of 200,000 MHz.* Only in 10 of the cases reported were two of the 1800
lines closer than 0.25 MHz.
There are two requirements that impose limitations on microwave studies.
(1) The spectrum must be obtained on the material in the gaseous state. For
conventional instrumentation, a vapor pressure of at least 10-' torr is required.
(2) The molecule must have a permanent dipole moment in the ground state in
order to absorb microwave radiation, since rotation alone cannot create a dipole
moment in a molecule.
In addition to the fingerprint application, other useful data can be obtained
from the microwave spectrum of a compound. Some of the most accurate bond
distance and bond angle data available have been obtained from these studies.
Let us first consider a diatomic molecule. The rotational energy, E, for a diatomic
molecule is given by the equation:
E = hBJ(J + 1) (6-19)
where J is an integer, the rotational quantum number; h is Planck's constant

(6.62 x 1027 erg sec); and B= h/8ir2 1 where I is the moment of inertia. Since
E = hv, we obtain the relationship for the frequency, v, corresponding to this
energy:
v = BJ(J + 1) (6-20)
The energy of the transition from state J to J + I can be determined by

substitution in equation (6-19), leading to
AE = 2Bh(J + 1)
or
Av = 2B(J + 1)
There is another selection rule (in addition to the dipole moment requirement)
for microwave absorption, which states that AJ = - 1. Therefore, the longest
wavelength (lowest energy) absorption band in the spectrum will correspond to
*A wave number of 1 mm ' is equivalent to a frequency of 299,800 MHz.

the transition from J = 0 to J = 1, for which the frequency of the absorbed energy
is:
2h
Av = 2B - (6-21)
Av is measured and all other quantities in equation (6-21) are known except I,
which can then be calculated.
All other lines in the spectrum will occur at shorter wavelengths and will
be separated from each other by 2B [i.e., AAv = 2B(J + 2) - 2B(J + 1) = 2B,
where (J + 1) is the quantum number for the higher rotational state and J that
for the lower; see Fig. 6-17]. Once the moment of inertia is determined, the
equilibrium internuclear separation, r,, in the diatomic molecule can be calculated
with equation (6-22):
I = pr. 2 (6-22)
where p is the reduced mass.
FIGURE 6-17 Illustration of the 0 -- 1

and other rotational transitions
2B 2B 2B
(idealized).
10'
In the above discussions we have assumed a rigid rotor as a model. This

means that the atom positions in the molecule are not influenced by the rotation.
Because of the influence of centrifugal force on the molecule during the rotation,
there will be a small amount of distortion. As a result, the bond distance (and
I) will be larger for high values of J, and the spacings between the peaks will
decrease slightly as J increases.
In a more complex molecule the moment of inertia is related to the bond
lengths and angles by a more complex relationship than (6-22). A whole series
of simultaneous equations has to be solved to determine all the structural
parameters. In order to get enough experimental observations to solve all these
equations, isotopic substitution is employed. For pyridine, the microwave spectra
of six isotopically substituted compounds (pyridine, 2-deuteropyridine, 3-
3
deuteropyridine, 4-deuteropyridine, pyridine-2- C, and pyridine-3- C) were
employed. For a more complete discussion of this problem, see Gordy, Smith,
and Trambarulo. 5)
By a detailed analysis of the microwave spectrum, one can obtain information
about internal motions in molecules. The torsional barrier frequency and energy
7
can be obtained." )
Double resonance experiments have also proved valuable for assigning peaks
and for determining relaxation times of states. A given transition is saturated
with microwave power, producing a non-equilibrium distribution in the popu-
lations of the two levels involved. One then looks for an increase or decrease in
other transitions that are thought to involve one of these levels.(17b)
6-17 Procedures 179
Accurate dipole moment measurements can be made from microwave
experiments. When an electric field is applied to the sample being studied, the
rotational line is split (the Stark effect). The magnitude of the splitting depends
upon the product of the dipole moment (to be evaluated) and the electric field
strength (which is known).
Much information is available from the Zeeman splitting of rotational
lines.(1 7 c'd) The components of the magnetic susceptibility anisotropy can be
evaluated, as can the expectation value Kr2 >. As we shall see in the chapter on
nuclear quadrupole resonance, nqr, much the same information obtained on
solids by nqr can be obtained on gases by microwave spectroscopy.
6-16 ROTATIONAL RAMAN SPECTRA
Information equivalent to that obtained in the microwave region can be obtained

from the rotational Raman spectrum, for which the permanent dipole selection
rule does not hold. As a result, very accurate data on homonuclear diatomic
molecules can be obtained from the rotational Raman spectrum. Experimentally,
the bands are detected as Stokes lines with frequencies corresponding to rotational
transitions.
In order for a molecule to exhibit a rotational Raman spectrum, the
polarizability perpendicular to the axis of rotation must be anisotropic; i.e., the
polarizability must be different in different directions in the plane perpendicular
to the axis. If the molecule has a threefold or higher axis of symmetry, the
polarizability will be the same in all directions and rotational modes about this
axis will be Raman inactive. Other rotations in the molecule may be Raman active.
For diatomic molecules the selection rule AJ = +2 now applies:
AE = Bh(4J' - 2) (6-23)
and the frequency separation between the lines is 4B. If this separation is not at
least ~0.1 cm -' the lines cannot be resolved in the Raman spectrum. The rest
of the calculation is identical to that described previously. There has been only
limited application of this technique to more complex molecules.
6-17 PROCEDURES Applications of

a. In Infrared
Infared and
Raman
The most common infrared equipment covers the wavelength region from 4000 Spectroscopy*
to 400 cm--. The grating resolves polychromatic radiation into monochromatic
radiation so that variations in absorption of a sample with change in wavelength
can be studied. The resulting spectrum is a plot of sample absorbance or percent
transmission versus wavelength (see Fig. 6-26).
* The reader is referred to the text by Nakamoto'( and to reference 11 for extensive compilations
of examples of infrared and Raman spectroscopy, respectively.
Movable Stationary The Fourier transform (FuR) technique, 191 which will be discussed in more
9
mirrormirrormirror
mro detail in Chapter 7 on nmr, is now very common in infrared instrumentation.(1 )
r,
r2 . Source spectrum. Thus, multiple scanning of the spectrum and computer storage of the
r2 ri scans become feasible, greatly enhancing the signal-to-noise ratio. Furthermore,

filtering is not required, eliminating the need for energy-wasting slits. Enhanced
Ar 0 Ar resolution in a shorter time period is possible. A schematic illustration of the
(A)method is shown in Fig. 6-18(A).
The monochromatic source is split into two components. One is sent to a
stationary mirror traveling a distance r. The second component is sent to a mirror
moves,
.thattraveling a distance of re ± Ar. The reflected components are
W/2 + refocused at the detector. When Ar 0,n re= r2 a, the beams reinforce. When
fI
the mirror is moved so that r2 rs 2 v p+ the
A/4,radiation travels r2 ave o+t2
(B) (/'.m4 going and coming back) and is 180 out of phase with the component that
traveled r. The beams cancel. If the mirror is moved continuously, the intensity,
sb' of the radiation at the detector varies as shown in Fig. 6-18(B). If the mirror
is moved with a constant velocity V,, over a distance 2ra2, one cycle of a sine
wave is produced.
ICffThe frequency at the detector, va, is related to the wavelength by
FIGURE 6-18 (A) Schematic
VD VM
Y cm (6-24)
of the FTIR technique. (B)
Radiation sensed at the
detector for a continuously With a wavelength 2 of 10 p (i.e., a frequency of 1000 cm -1) and a mirror velocity
moving mirror. (C) Inter- of 0.5 mm/sec-1, a frequency of 50 Hz (cps) results. The low frequency signal
ferrogram from a that results from the high frequency source is termed an interferogram. Poly-
polychromatic source. chromatic radiation is employed in FTIR. Each component produces a wave
with a unique frequency. The summation of all such waves produces an
interferogram as shown in Fig. 6-18(C). All frequencies add only at r, = r2 . The
Fourier transform of the interferogram converts the time axis to a frequency axis
and gives all the frequency components of the source along with their intensities.
If the beam passes through a sample before hitting the detector, some frequencies
will be absorbed and missing in the Fourier transform. The spectrum is obtained
by difference. A laser is used as a reference source to constantly calibrate the
frequency scale of the instrument.
Infrared spectra can be measured on gases, liquids, or solids. Liquids and
gases are generally studied by absorbance measurements. Certain wavelengths
are absorbed by the sample and the others are transmitted. The materials used
to construct the cells must not absorb in the regions of interest. Special cells with
a long path length are needed for most gaseous samples.
Solutions are commonly studied. The solvent absorption must be subtracted
out. Different techniques are used depending on the instrumentation. In regions
where the solvent absorption is very large almost all of the radiation is absorbed
by the solvent and the sample and no sample absorptions can be detected in
these blank regions. The open regions for several common liquids used for solution
work are illustrated in Fig. 6-19.
Absorbance spectra of solid samples are often examined as mulls. The mull
is prepared by first grinding the sample to a fine particle size and then adding
enough oil or mulling agent to make a paste. The paste is examined as a thin
layer between sodium chloride (or other optical material) plates. The quality of
the spectrum obtained is very much dependent on the mulling technique. When
6-17 Procedures 181
Wavelength [gm]
3 41 5 6 7 8 9 10 12 1416 20 30
11 c 1 I 11 111 i 1 1 1 1 1 I I I I 1 1111111b)il l I 11110 1 hhlh1 mii
Acetone (b)
Acetonitrile (b)
Benzene (b)
Chloroform (b)
Diethylether (c)
Dichloromethane (b)
N, N - Dimethylformic-
amide (c)
Dimethylsulfoxide (c)
Dioxane (c)
n - Hexane (b)
Paraffin (Nujol) (b)
Poly (chlorotrifluoroethane) (a)
Pyridine (b)
Carbondisulfide (a)
Tetrachloroethene (b)
Carbontetrachloride (a)
Tetrahydrofuran (c)
Toluene (b) 3500 3000 2500 2000 1800 1600 1400 1200 1000 800 600 400
Wavelength (cm-')
FIGURE 6-19 Regions in which various solvents transmit less than 20% of the
incident radiation (indicated by solid bars). (A) 200 pm cell thickness; (B) 100 pm;
(C) 20 pm. See references 18 and 19 for other solvents.
the spectrum is examined, peaks from the mulling material will appear in the
spectrum and possibly mask sample peaks. If two spectra are obtained, one in
Nujol and one in hexachlorobutadiene, all wavelengths in the 5000 to 650 cm -1
region can be examined. Solid samples are sometimes examined as KBr discs.
The sample and KBr are intimately mixed, ground, and pressed into a clear disc
that is mounted and examined directly. Care should be exercised in this procedure,
for anion exchange and other reactions may occur with the bromide ion during
grinding or pressing.
Transmission IR can be used1201 to study surfaces of solids that have
transparent windows in the infrared, e.g., oxides of silica and alumina. High
surface area materials should be used so that the spectra have an appreciable
contribution from the surface as opposed to the bulk. These samples can be
pressed into discs and inserted into the beam path.
Reflectance techniques 21 ) can be applied to samples that cannot be studied
by transmission. These methods can be classified into general categories-
specular and diffuse reflectance. Specular reflectance 22 ) is the front surface,
mirrorlike reflection off a smooth surface such as a metal foil or single crystal
surface. Diffuse reflectance is observed with rough surfaces and powders. Since
the incident radiation is reflected diffusely, it must be collected and refocused on
a detector. The reader is referred to reference 22 for the factors that must be
considered in comparing absorption and reflection spectra.
Source Pressure An additional technique, referred to as IR-External Reflection spectros-

copy,( 2 ) involves studying films coated on highly reflecting surfaces, e.g., gold.
Large angles of incidence enhance the sensitivity of the technique. A polarization
dependence of the vibrational modes of the adsorbed molecule provides orientational
I, information about the adsorbed molecule when polarized light is employed.
Photoacoustic Fourier transform spectroscopy (PAS)"23 provides an addi-
tional technique for obtaining the infrared spectrum of surface molecules.
Absorption of light energy by a confined gas heats it and increases its pressure.
Sample Microphone If the light source is modulated there will be a modulation of the gas pressure.
Sound is generated at the frequency of the light modulation and can be detected
FIGURE 6-20 Schematic with a microphone. Modulated light absorbed by a solid heats its surface, which
illustration of the in turn conductively heats a boundary layer of air next to the surface as shown
photoacoustic technique. in the Fig. 6-20. Sound is again generated, but the signal detected is reduced in
amplitude and delayed in phase because the heat must be conductively transmitted
to the surface. For solids, only the energy absorbed near the sample surface is
effective in heating the surface quickly enough to contribute to the photoacoustic
signal. Thus, PAS is a true surface technique. In the FTIR/PAS experiment, each
wave number of light is modulated at a unique frequency and this frequency
information is preserved in the microphone signal. Thus, PA detection is a sort
of thermal detection in which the light absorbed heats up the sample. PA spectra
are qualitatively and quantitatively similar to transmission spectra and are
commonly ratioed to a saturated reference (carbon black). Sample surface
morphology has a minor qualitative effect on the spectra. Depending on the
source employed, one can do PAS in the IR, near IR, u.v., or visible. Diffuse
reflectance infrared Fourier transform spectroscopy (DRIFTS), is com-
pared(2 ' 2 1 ) with PAS to enable one to choose the best method for the problem
at hand.
The spectrum obtained for a given sample depends upon the physical state
of the sample. Gaseous samples usually exhibit rotational fine structure. This
fine structure is damped in solution spectra because collisions of molecules in
the condensed phase occur before a rotation is completed. In addition to the
difference in resolved fine structure, the number of absorption bands and the
frequencies of the vibrations vary in the different states. As an illustration, the
fundamentals of SO 2 are reported in Table 6-7 for the different physical states.
Other effects not illustrated by the data in Table 6-7 are often encountered.
Often there are more bands in the liquid state than in the gaseous state of a
substance. The stronger intermolecular forces that exist in the solid and liquid
states compared with those in the gaseous state are the cause of slight shifts
TABLE 6-7. Frequencies (in cm -1) for SO 2

Fundamentals in Different Physical States
Gas Liquid' Solid
v2(ai) 518 525 528

v1 (ai) 1151 1144 1144
v3 (bI) 1362 1336 1322
1310
'Raman spectrum.
6-17 Procedures 183
(assuming that there are no pronounced structural changes with change of phase)
in frequencies. Frequently, new bands below 300 cm-' appear in the solid state
spectrum. The causes of these frequency shifts, band splittings, and new bands
in the condensed states are well understood12" and will be discussed briefly here.
Bands below 300 cm-' are often caused by lattice vibrations in solids, i.e.,
translational and torsional motions of the molecules in the lattice. These are
called phonon modes and are responsible for the infrared cutoff of NaCl cells at
low wave numbers. Heavier atom alkali halides have phonon modes at lower
energies. These vibrations can form combination bands with the intramolecular
vibrations and cause pronounced frequency shifts in higher frequency regions of
the spectrum. An additional complication arises if the unit cell of the crystal
contains more than one chemically equivalent molecule. When this is the case,
the vibrations in the individual molecules can couple with each other. This
intermolecular coupling can give rise to frequency shifts and band splitting.
As mentioned earlier, molecular symmetry is very important in determining
the infrared activity and degeneracy of a molecular vibration. When a molecule
is present in a crystal, the symmetry of the surroundings of the molecule in the
unit cell, the so-called site symmetry, determines the selection rules. Often bands
forbidden in the gaseous state or in solution appear in the solid, and degenerate
vibrations in the gaseous state are split in the solid. The general problem of the
effect of site symmetry on the selection rules has been treated theoretically.(26 >
As a simple illustration of this effect, the infrared spectra of materials containing
carbonate ion will be considered. The infrared and Raman spectra of CaCO, in
calcite, where the carbonate ion is in a site of D3 symmetry, contain the following
bands (in cm '): vi, 1087 (R); v2 , 879 (IR); v3, 1432 (IR, R); v4 , 710 (IR, R). The
(IR) indicates infrared activity and the (R) Raman activity. The infrared spectra
of CaCO 3 in aragonite, where the site symmetry for carbonate is C, differs in
that v, becomes infrared active and v3 and v4 each split into two bands. By using
the symmetry considerations previously discussed on the CO 32 ion, the following
results can be obtained:
V1 V2 V3 V4
D3 h Ai'(R) A 2"(IR) E'(IR, R) E'(IR, R)

D3 A 1(R) A 2(IR) E(IR, R) E(IR, R)
C, A'(IR, R) A"(IR, R) A'(IR, R) + A'(IR, R) A'(IR, R) + A'(IR, R)
As a result of all these possible complications, the interpretation of spectra

obtained on solids is difficult.
The spectrum of a given solute often varies in different solvents. In
hydrogen-bonding solvents the shifts are, in part, due to specific solute-solvent
interactions that cause changes in the electron distribution in the solute. The
frequency of the band is also dependent upon the refractive index of the solvent(' 7 )
and other effects.(2 8 )
Matrix isolation experiments combined with infrared and Raman studies
have led to interesting developments. Unstable compounds, radicals, and inter-
mediates are trapped in an inert or reactive solid matrix by co-condensing the
matrix (e.g., argon) and the species to be studied at low temperatures (often2 9 4.2
to 20 K). Uranium, platinum, and palladium carbonyls have been prepared( ab)
by allowing controlled diffusion of CO into an argon matrix of the metal. The
compounds LiO 2 , NaO 2 , KO2 , RbO 2 , and LiN 2 have also been made( 29c) and
investigated by infrared. The findings are consistent with C 2, symmetry. 2Several
9
interesting species containing bound 02, CO, and N2 have been made.( d)
b. In Raman
Raman spectra are routinely run on gases, liquids, solutions, or pressed pellets.
Liquids can be purified for spectral studies by distillation into a sample cell.
Removal of dust is essential, since such contamination increases the background
near the exciting line. Turbid solutions are to be avoided. Water and D 2 0 are
excellent Raman solvents because, in contrast to their behavior in the infrared,
they have good transparency in the Raman vibrational region. Figure 6-21
illustrates the cutoff region for various solvents in the Raman region. A complete
Raman spectrum can be obtained by using CS 2 , CHCl3 , and C 2 C14 . For colored
solutions, an exciting line should be used that is not absorbed. Fluorescence and
photolysis are problems when the sample absorbs the exciting line.
Solids are best investigated as pellets or mulls. The techniques for preparing
these are the same as those for infrared. If powders are examined, the larger the
crystals the better. For colored materials, rotating the sample during the
experiment prevents sample decomposition and permits one to obtain a good
spectrum. Surface scanning techniques are also helpful.(9). Experimental aspects
are described in more detail in references 10 and 11.
6-18 FINGERPRINTING
If an unknown is suspected to be a known compound, its spectrum can be

compared directly with that of the known. The more bands the sample contains,
the more reliable the comparison.
The presence of water in a sample can be detected by its two characteristic
absorption bands in the 3600 to 3200 cm 'region and in the 1650 cm region.
C2 ir a amae
CC R- a m
Cd 4 r O
D am
CHC13 r c o
CH2C2 R a I
C 22
C l
ir
ia e U. gU
C2C1 mm i3
4
HCCN R U
H 20
020 C3
3000 2000 1000

WAVENUMBERS - cm 1
FIGURE 6-21 Solvent interference regions in the Raman region. Solid regions
are completely obscured, open rectangles partially obscured.
If the water is present as lattice water, these two bands and one in the 600 to
300 cm-' region are observed. If the water is coordinated to a metal ion, an
additional band in the 880 to 650 cm-1 region is often observed."g1 In the
clathrate compound Ni(CN) 2 NH 3 -CH 6 -xH 2 O, the infrared spectrum(2 3b) clearly
showed the presence of water, although it was not detected in a single crystal
x-ray study. In connection with the earlier discussion on the effect of physical
state on the spectra of various substances, it is of interest to mention in passing
that the benzene is so located in the crystal lattice of this clathrate that the
frequencies of the out-of-plane vibrations are increased over those in the free
molecule, while the in-plane vibrations are not affected.
If the product of a reaction is suspected of being contaminated by starting
material, this can be confirmed by the presence of a band in the product due to
the starting material that is known to be absent in the pure product. Even if
spectra are not known, procedures aimed at purification, e.g., recrystallization,
may be attempted. Changes in relative band intensities may then indicate that
partial purification has been achieved. Consistent spectra on repeated purification
steps may be used as a check on completeness of purification.
A more common application employs the group vibration concept to
ascertain the presence or absence of various functional groups in the molecule.
The following generalizations aid in this application:
1. Above 2500 cm-- nearly all fundamental vibrations involve a hydrogen
stretching mode. The O-H stretching vibration occurs around 3600 cm-1.
Hydrogen bonding lowers the frequency and broadens the band. The N-H
stretch occurs in the 3300 to 3400 cm-' region. These bands often overlap the
O-H bands, but the N-H peaks are usually sharper. The N-H stretch in
ammonium and alkylammonium ions occurs at lower frequencies (2900 to 3200
cm -1). The C-H stretch occurs in the 2850 to 3000 cm- 'region for an aliphatic
compound and in the 3000 to 3100 cm-1 region for an aromatic compound.
Absorptions corresponding to S-H, P-H, and Si-H occur around 2500, 2400,
and 2300 cm-1, respectively.
2. The 2500 to 2000 cm-1 region involves the stretching vibration for triply
bonded molecules. The C=N group gives rise to a strong, sharp absorption that
occurs in the 2200 to 2300 cm-1 region.
3. The 2000 to 1600 cm-1 region contains stretching vibrations for doubly
bonded molecules and bending vibrations for the O-H, C-H, and N-H
groups. The carbonyl group in a ketone absorbs around 1700 cm '.Conjugation
O O
in an amide RC-N(CH 3)2 <--RC=N(CH 3)2 decreases the C-0 force

constant and lowers the highly coupled (with C-N) carbonyl absorption to the
1650 cm-' region. Hydrogen bonding lowers the carbonyl vibration frequency.
Stretching vibrations from C=C and C=N occur in this region.
4. The region below 1600 cm-' is referred to as the fingerprint region for
many organic compounds. In this region significant differences occur in the
spectra of substances that are very much alike. This is the single bond region
and, as mentioned in the section on coupling, it is very common to get coupling
of individual single bonds that have similar force constants and connect similar
masses (e.g., C-0, C-C, and C-N stretches often couple). The absorption
BO_
oCO i
Co --
cs
(CO) co)
e(
2 N(NO) SiSiOdWJ
oNO t (PO) , (oP
PO
> (so) 1toso)
(SO) ( 0)
C10
(.(OCcO)
o BrO Br
FIGURE 6-22(A), (B), and
(C) Range of infrared group BX
frequencies for some i- CI--
inorganic and organic AIX A-- F- C g 4-
---
materials. The symbol vmx,,
CX c--e
where M and X are general cxci--F
symbols for the atoms Six
SiX
involved, corresponds to a
stretching vibration. The GeX o GX ~F - I~4Br
symbols vj, v2, etc, where
previously defined. The NX _cIII
*i -_-F-I
symbol 6 corresponds to an . PX
in-plane bending vibration (6,
is a symmetric bend, bd is an AsX
asymmetric bend), 7r to an
out-of-plane bend, p to ox '- r-4
rocking and wagging SX -c
vibrations. I-s-
SeX IF
2400 2000 1500 1000 500
i (cm-')
(A)
bands in this region for a given functional group occur at different frequencies
depending upon the skeleton of the molecule, because each vibration often involves
oscillation of a considerable number of atoms of the molecular skeleton.
Many characteristic infrared group frequencies have been compiled to aid
in this application. Figure 6-22 shows several typical plots.
In the far region, it becomes very difficult to assign the observed spectrum
to group frequencies. Figure 6-23 is a compilation that illustrates typical regions
in which absorption bands are found for molecules containing particular
functional groups.
6-19 SPECTRA OF GASES
The spectra of gases are often very much different from the spectra of materials
in the condensed phase or in solution. As can be seen in Fig. 6-20, a significant
difference is the presence of considerable fine structure in the gaseous spectrum.
The fine structure is due to a combination of vibrational and rotational transitions.
For example, in a diatomic molecule, there are not only transitions corresponding
to the pure vibrational mode, v0, but also absorption corresponding to v, + v,
1/- V/3
NCO- F-- H
'i, V/3
N3
NCS-
N02 H H
- %,,'V3
C032- ," i
H H
B0 33 '~'
0 2 +2 -v v4
C10 3
H HH FIGURE 6-22 (continued)
IO3 -II H-
SO42- "*
v-2
C104-
V2
PO4 3-
V/3'/I v4
MnO 4 V3 V H
V4 V2
CrO 42 ~ F-4
/ 1 '
SeO42- 4000 3000 2000 1500 1000 800 60C
2 ,v 3 4 v2 I I I I I I I
MOO 4 - H' 'C-H aC-H 'C-H
3 !H 4 V2 8
AsO4 ~ 'H VC-D C-o rC-O
|/ 2|
8
2 v0-H 0-H
Wa4 -
w(MC) 8(CMC) '/N-H BN-H
M-CN '-
v(CS) 8(NCS) /C!C "C-C "C-C
M-NCS*
v(CO) "C2N -C-N 'C-N
v(MC) H- " i-- I-
M-CO* VC O VC-O
a(NH ) 8,(NH) p,(NH 3 ) 1 (MN) 8 (NMN)
M-NH 3 VNO (C-N=O orO-N-0)
-aE
) (MN)
*o 0 M-NH 2 (For Hg 2complex)SQ H2 ) pjNH2 ),pt(NH 2 ) pr(NH 2 SNO2(C-NO 2 ,-ON0 2 )
v(O) N p.(NO2) (MN)
4)o M-N0 2* Mb H t- i -
H H
(ONO) v(ONO) 8(ONO) 'id C02'sco 2
-Ec5' M-ONO H H H
v(CC ) v(CO) v(CO) 7r (OCO) v(MO)
M-OC02" - -' F-- "I "F-1 I'C-F Uc.Cl
v NO) v(NO) v(NO) -r
M-ON0 FI H H UP.O
2
V(SO) S(OSO) Bd(OSO)
* M-OS0 3* VP-H
M(HOH) p,(OH2 ), p,(OH 2), p,(OH2 )
M-OH 2 H ' | "S-H
' I| H
2000 1500 1000 500

5(cm~') S I I
4000 3000 2000 1500 (000 800 600
1
v (cm- )
(C)
800cr 600 c m 400 c m 200 cm
hrancha
Alies, srihtCR
cre i ctrans gauche

.C C mmm.
hii
Alkuenes,RCH=CRR .
trans RCH=C
cis RCH=cCHR
RRC=CH .--
-CCI,
Car* -----
RCHCH= CH,-
=CFCI
=CF1r
CFri
Allenes m.
Alkynes R-C=CH
alky-CCH
CC-CI
..
aromatics I
1,2
Amrdes
1.3
1,4
Far infrared "24 -
FIGURE 6-23
vibrational frequency 1,a2,3

1,2,3,4 -ms
correlation chart.
m.
-.
1,23,4,5 m
Polyphenylis on
ortho== =
meta .. ...
#l orr
AAC I.
#=
Heteroyclics thiophones ..
m-dioxane
"pnMy
Nitriies
Nitr-
Amines NH,
Amides
Esters acetates -
Acids..
Ketones m =| m
Phosphorus
-organo (RriP-
(RO),POH-
(RO) POR-
(R0)hPO
(RO),(PS)SM
(RO):PSR -
(RO) PS - - -y
Sulfur -organo R-S-.R-
CH,-S- -__.-
R-CHS.
RRCH-S-=
RRRC-S.-
R-S-(C==0)-S#-R
R-(C=0)-S #
-(C=)-S- ..
Silieonrg o Si-,
SiHI
SO _ _i_-
SiO
SiS I
sip
SiSi
alkyl-S.- -. I- -
phenyi.Si_
Metal organo, phenyl-Ge - I-
phenyl-Sn
phenlPb-
alkyl Ge - -
alkylSn -..
alkyl-Pb
etc.-
Ferrocene. - I-
P thalocyanines
Aretylacetonates 1 -i
- -
Oxalato 4 -I- ±______ .L ~ 9
Inorgani
cs cyanides -t-tr+T
carbonyls 1.----- -
azides
PCI _
-
-t--- Im
PBr2________
Ph
-. 1 Cl. FIGURE 6-23 (continued)
-i
-SF,
S.
PD.
COO.
MnS,
RaS
Moo,
WO.
TCn,
10, i
sio,
_ '
800 600 400 200
WAVENUMBER IN cm-1
where v, represents the rotational frequency detected. Since any finite sample
contains many molecules, many different rotational states will be populated and
there will be a whole series of lines corresponding to different v, values (i.e.,
transitions between many different rotational states). This phenomenon is
illustrated in Fig. 6-24. The Q-branch corresponds to the transition in which v,
is zero (i.e., a transition with no change in rotational quantum number), the
R-branch to v 0 + v,, and the P-branch to v0 - v,.
The frequencies of all the bands in Fig. 6-24 can be expressed by the following
equation:
2
v = v, + hm,/87r2 1
where mj has all integral values, including zero, from +J to -J (where J is the
rotational quantum number) depending on the selection rules. When mj = 0, the
Q branch
FIGURE 6-24 Schematic of the R branch P branch
transitions giving rise to P, Q,
and R branches in a spectrum
of a gas.
vibrational transition is one that occurs with no change in rotational quantum

number. A Q-branch results. When m, is less than zero (J ,, -- J, transitions),
lines in the P-branch result, whereas lines in the R-branch correspond to mj
greater than zero.
A series of selection rules for the combination of vibration and rotation
transitions in various molecules is helpful in deducing structure:
1. Diatomic Molecules. Most diatomic molecules do not possess a

Q-branch. Nitric oxide (NO) is the only known example of a stable diatomic
molecule that has a Q-branch. The diatomic molecule must possess angular
momentum about the molecular axis in order to have a Q-branch. (Y states,
L = 0, have no Q-branch.)
2. Linear Polyatomic Molecules. If the changing dipole moment for a given

vibrational mode is parallel to the principal rotation axis in the molecule, a
so-called parallel band results, which has no Q-branch. The selection rule for this
case is AJ = + 1; AJ cannot be zero. If the dipole moment change for the vibration
has any vector component perpendicular to the principal axis, a perpendicular
band will result with a Q-branch, and AJ can be 0, + 1. The asymmetric C-O
stretch in CO 2 is a parallel band, while the C-O bending vibration is a
perpendicular band. The utilization of these criteria in making and substantiating
assignments of vibrations is apparent. If, in the spectrum of a triatomic molecule,
any of the infrared bands (vi, v2 , or v3 ) consists of a single P- and R-branch
without any Q-branch (i.e., a zero-gap separation), the molecule must be linear.
This type of evidence can be used to support linear structures for N20, HCN,
and CO 2 . The rotational spacings can be employed, as discussed under the section
on microwave spectroscopy, to evaluate the moment of inertia. Note that even
though CO 2 does not have a permanent dipole (and hence is microwave inactive),
the rotational spacings can be obtained from the fine structure in the infrared
spectrum.
3. Non-linear Polyatomic Molecules. For the discussion of vibration-

rotation coupling in non-linear polyatomic molecules it is necessary to define
the terms spherical, symmetric, and asymmetric top. Every non-linear molecule
has three finite moments of inertia. In a spherical top, e.g., CCl 4 , all three moments
are equal. In a symmetric top, two of the three moments are equal. For example,
of one selects the C 3-axis of CH 3 Br as the z-axis, then the moment of inertia
along the x-axis is equal to that along the y-axis. Any molecule with a threefold
or higher rotation axis is a symmetric top molecule (unless it is a spherical top).
In an asymmetric top, no two of the three moments are equal. Molecules with
no symmetry or those with only a twofold rotation axis (H 2 O, for example) are
asymmetric top molecules.
C
0
4)4
I I I i
1340 1320 1300 1280 1260
S(cm-)
FIGURE 6-25 A parallel band for CH3Br (symmetric top).

[From G. M. Barrow, "Introduction to Molecular
Spectroscopy". Copyright @ 1962 by McGraw-Hill, Inc. Used by
permission of McGraw-Hill Book Company.]
For symmetric top molecules, those vibrations having oscillating dipole

moments that are parallel to the principal axis produce a parallel absorptionband
with P-, Q-, and R-branches. The symmetric C-H stretching and bending
frequencies in CH 3 Br are examples of parallel bands. The type of spectrum
obtained for a parallel band is illustrated in Fig. 6-25. In this example the
rotational fine structure in the R-branch is not resolved. The parallel band for
a symmetric top molecule is similar to the perpendicular band for a linear
molecule. For a perpendicular absorption band in a symmetric top molecule,
several Q-bands are detected, often with overlapping unresolved P- and R-
branches. The C-Cl bending vibration in CH 3 Cl is an example of a perpendicular
band in a symmetric top molecule. A typical spectrum for this case is illustrated
in Fig. 6-26. In a spherical top, the selection rule for a perpendicular band is
AJ = 0, + . This information can be employed to support band assignments.
0
V
U)
CQ
Q
i (cm')
FIGURE 6-26 Perpendicular band containing two Q branches.
F F F F
N S N S
S N S N
F F F F F F F F
FIGURE 6-27 Some possible C2v C'

structures for NSF 3. N N S
F-S-F S N
F F F F
F FF
(Planar)
C2, C3, C 3,
It becomes very difficult in some cases to distinguish between symmetric

and asymmetric top molecules. As the symmetry of the asymmetric top becomes
lower, the task becomes easier. Considerable information can be obtained from
the band shape to support vibrational assignments.*
6-20 APPLICATION OF RAMAN AND INFRARED

SELECTION RULES TO THE DETERMINATION OF
INORGANIC STRUCTURES
1. The structures in Fig. 6-27 represent some of the possibilities that should
be considered for a material with an empirical formula NSF 3 . Table 6-8
summarizes the calculated number and symmetry of bands for these possible
structures. These results and the infrared activity of the bands should be
determined for practice, employing the procedure in the section on symmetry
considerations and the character tables in Appendix A.
The infrared spectrum has been reported. Six intense bands are found,
in agreement with the C3, structure but certainly not in conflict with other
possible structures. Some of the fundamentals may be of such low intensity that
they are not seen. However, it was found that four of the bands (v , v , v1 3)
1 2
have P-, Q-, and R-branches. This is good evidence that the molecule being
investigated is a symmetric top molecule and supports a C3, structure. The
spectral data are contained in Table 6-9, where they are compared with data
reported for POF3. (30b) The forms of the vibrations are diagrammed in Appendix
C (see the discussion of the ZXY 3-molecule). These data, combined with an nmr
study, have been employed to support the C 3 . structure F 3S-N. Polarized
Raman studies were not carried out. These would not add a great deal of structural
information but would help confirm the assignments. The A, bands should be
polarized and the E bands depolarized.
The F 3N-S structure has not been eliminated by these data. The frequencies
of the observed bands provide information to favor the F 3SN structure. The
* See Herzberg (in Additional References), pp. 380-390, for an explanation of the phenomena
just discussed. Many applications of these principles are discussed in this reference.
6-20 Application of Raman and Infrared Selection Rules to the Determination of Inorganic Structures 193
TABLE 6-8. Calculated Number of Bands for Various

Structures of NSF 3
C2,
C", (C 2 -axis is the z-axis) C,
3A, (IR active) 4A 1 (IR active) 7A' (IR active)

3B, (IR active) 2A" (IR active)
3E (IR active) 2B 2 (IR active)
S-N stretch is the only fundamental vibration that could be expected at a TABLE 6-9. Fundamental
frequency of 1515 cm- '. [A very rough approximation for this frequency can be Vibration Frequencies for
obtained by employing the Cz=C force constant and the S and N masses in NSF 3 and OPF 3
equation (6-3).] The NF stretching vibrations in NF 3 occur at 1031 cm-' and
NSF 3 OPF 3
would not be expected to be as high as 1515cm-' in F 3N-S. The FSN structure
receives further support from microwave and mass spectroscopic studies. The v1 1515 1415
low moment of inertia calculated from the microwave studies cannot be explained v2 775 873
unless the sulfur atom is at the center of mass. v3 521 473
811 990
2. The same approach as that outlined above was employed to prove that
v1 429 485
perchlorylfluoride, ClO 3F, has a C 3 , structure instead of O2 CIOF.0 1c) 6 342 345
3. An appreciation for some of the difficulties encountered in making
31
spectral assignments can be obtained from an article by Wilson and Hunt a)
on the reassignment of the spectrum of SO 2 F2 . Often a set of band assignments
may appear self-consistent, but these assignments may not be a unique set. The
effect of isotopic substitution on spectral shifts can be employed to provide
additional support for the assignments made. A Raman study has also been
3
reported.( "1)
4. The infrared and Raman spectra have been interpreted to indicate a
planar structure for N 2 O4 .P) Structures corresponding to the planar D2 model
and the staggered D2 d configuration were considered. Since the infrared and
Raman spectra have no lines in common, it is proposed that the molecule has a
center of symmetry. A rigorous assignment of the bands could not be made with
the available data.
32
5. The spectral data for B 2H, support the bridged hydrogen structure.( a)
The entire spectrum is analyzed, assignments are made, and conclusions regarding
the structure are drawn from the rotational fine structure of some of the
perpendicular bands.
6. An infrared and Raman study(3 2 b) supports a C 2 , planar T-shaped
structure for CIF 3 . Too many fundamentals were observed for either a D3h or
C 3, structure. Some of the bands had P-, Q-, and R-branches, while others had
only P- and R-branches. Since all fundamentals for a tetraatomic C 3, molecule
(e.g., PF 3) have P-, Q-, and R-branches, this structure is eliminated.
7. The infrared and Raman spectra of B 2(OCH 3)4 have been interpreted
to indicate(3 2 c a planar arrangement for the boron and oxygen atoms. Many
bands were found in the spectrum. Assignments were made that are consistent
with a planar C2 molecule. More fundamentals were found than would be
predicted from a D2 model. The infrared and Raman spectra do not have a
coincidence of any of the vibrations asigned to B-B or B-O modes; but, as
would be expected for the proposed structure, the CH, and O-C vibrations are
found in both the infrared and Raman. Good agreement is found for the bands
assigned to the O-C and -CH 3 modes in this molecule and in the B(OCH 3)3
molecule, illustrating the applicability of the group frequency concept. The above
assignments indicate a planar structure, but since the interpretation is very much
dependent on the correctness of the assignments, they should be confirmed by
isotopic substitution. The band assignments made are consistent with those
calculated from a normal coordinate analysis assuming a C2, model.
One final advantage of combining Raman studies with infrared will be
mentioned. Often, in the infrared spectra, it is difficult to distinguish combination
bands or overtones from fundamentals. This is usually not a problem in the
Raman, for the fundamentals are much more intense. An extensive review of the
inorganic applications of Raman spectroscopy' ) should be consulted for other
applications of this technique.
Bond Strength There are many examples in the literature of attempts to infer the bond strength
for coordination of a Lewis acid to Lewis bases from the magnitude of the infrared
Frequency Shift shift of some acid (or base) functional group upon coordination. Usually a
Relations relationship between these two quantities is tacitly assumed, and the infrared
shift interpreted as though it were a bond strength in terms of the electronic
structures of the acid and base. This approach is to be discouraged, for the
problem is too complex to permit such an assumption.
The systems that have been most thoroughly studied are the hydrogen
bonding ones. In general, hydrogen bonding(") to an X-H molecule results in
a decrease in the frequency and a broadening of the absorption band that is
assigned to the X-H stretching vibration. The spectra of free phenol, A (where
X = CHO-), and a hydrogen-bonded phenol, B, are indicated in Fig. 6-28.
The magnitude of AVOH, the frequency shift upon formation of a 1:1 complex
with a base, is related to - AH for the formation of the 1:1 adduct measured in
a poorly solvating solvent, for a large series of different bases as shown1341 in
Fig. 6-29. Different alcohols require different lines, as illustrated by the separate
A Free phenol in C Cl4

B Phenol plus donor in CC14
FIGURE 6-28 Infrared

A-
spectra of pehnol and a
hydrogen-bonded phenol.
I I I I
3600 3500 3400 3300 3200
-v (cm~')
6-20 Application of Raman and Infrared Selection Rules to the Determination of Inorganic Structures 195
04 4 1 PHENOLS
EtOAc
8.0-
2-
6.0-
t-BUTANOL
4.0
0.0
100 200 300 400 500 600 700 800 900
APOH(CM-1)
FIGURE 6-29 Constant-acid.constant-base frequency shift-enthalpy relations. Solid

lines are constant-acid lines. The phenol line also contains values for
p-t-butylphenol (+), phenol (0), p-chlorophenol (ID), and m-trifluoromethylphenol
(A). The acid butanol is represented by 0 and (CF3)2CHOH by A. The number 1
represents all points on the pyridine constant-base line; 2, ethyl acetate; and 4,
N,N-dimethylacetamide. (The pyridine-(CF 3)2CHOH Av is estimated from the Av vs.
Av plot of this acid and phenol.)
solid lines in this figure. The solid lines are referred to as constant-acid lines. The
dotted lines represent data in which the base is held constant and the acid is
varied. Over this range of enthalpies, a linear constant-acid equation exists for
various alcohols:
- AH(kcal mole -1) = (0.0 105 + 0.0007) AvOH + 3.0( + 0.2)(phenols) (6-25a)
-AH(kcal mole -') = (0.0115 + 0.0008) AvOH + 3.6(±0.03)[(CF3) 2CHOH]
(6-25b)
-AH(kcal mole- 1) = (0.0106 + 0.0005) AvOH + 1.65(±0.09)(t-butanol) (6-25c)
-AH(kcal mole-1) = (0.0123 + 0.0006) AvOH + 1.8(±0.1)(pyrrle 3 4)) (6-25d)
The authors(3 4 -38) caution against the use of this relationship for a new class of
donors (i.e., a functional group not yet studied) without first evaluating -AH
and AvOH for at least one system involving this functional group. The enthalpy
35
for a given acid has been shown to be related to two different properties( ) of
the base via the following relation:
-AH = EAEB + CACB (6 26)

where EA and EB are roughly related to the tendency of the acid and base to
undergo electrostatic interaction and CA and C, to their tendency to undergo
covalent interaction. Parameters for many different acids and bases have been
reported. When the range of C,/E, is between 1.5 and 6.0, the enthalpy-frequency
shift correlations hold to within 0.3 kcal mole '. For sulfur donors, where this
ratio is ~ 20, the plots of Fig. 6-29 do not hold.(3 6 >Unless AvoH should have the
same functional dependence on E, and C, as -AH does, there will have to be
a limitation on the C/E ratio of the base that can be used. Furthermore, there
is an experimental point at which AH and AvOH are both equal to zero.
Accordingly, the line will have to curve; this curvature is predicted by theory. 7 )
The theoretical treatment also predicts curvature on the high frequency end of
the curve. As a result, extrapolation of the curve in either direction is dangerous.
The analysis ( 7 ) indicates that frequency shifts can be employed instead of force
constants because the high-energy O-H stretching vibration is far enough
removed in energy from other vibrations that the band assigned to it is a nearly
pure O-H stretching vibration.
In view of the considerable effort involved in an enthalpy measurement (often
one or two days of work), this relationship is a welcome one for ascertaining the
strength of the hydrogen bond. However, much confusion has been generated
by the inappropriate use of this relation. Inappropriate solvent selection, and
enthalpies that are a composite of things other than the simple 1:1 adduct
formation reaction (e.g., two different donor sites in the same molecule(3 1)), should
be avoided.
Additional correlations of this sort involving different types of acids, and
others involving some spectroscopic change in the base upon coordination, would
be most useful. However, they cannot be naively assumed but must be tested
against measured enthalpies and the limitations sought. In a study of the Lewis
acid chloroform,(") it was found that the enthalpy of adduct formation did not
correlate with the observed frequency shift.
6-21 CHANGES IN THE SPECTRA OF DONOR

MOLECULES UPON COORDINATION
The infrared spectrum of N,N-dimethylacetamide in the solvent CC14 has an

absorption band at 1662 cm - ' that is due to a highly coupled carbonyl absorption.
The low frequency compared to acetone (1715 cm-1) is attributed to a resonance
interaction with the lone pair on the nitrogen (see Fig. 6-30). Upon complexation
with several Lewis acids, a decrease in the frequency of this band is observed.(40 )
This decrease has been attributed to the effect of oxygen coordination to the
acid. Oxygen coordination could have several effects upon the vibration:
O IO I
FIGURE 6-30 Resonance structures for C OH3 C CH3
N,N-dimethylacetamide. H3 C N <-> H3 C N
OH3 CH3
6-21 Changes in the Spectra of Donor Molecules upon Coordination 197
1. Since the oxygen atom has to move against the atom to which it is
coordinated, an increase in frequency should result. In effect this is to say that
for the system X-O=C , the C-O and X-O vibrations couple, producing
a higher-energy carbonyl absorption.
2. A change in oxygen hybridization could increase (or decrease) the C-O
* bond strength and increase (or decrease) the force constant of the C-O bond.
3. The most important effect in this case involved decreasing the carbonyl
force constant by draining 7n electron density out of the carbonyl group. This
causes the observed decrease in the carbonyl frequency and indicates oxygen
coordination. The absence of any absorption in the carbonyl region on the
high-frequency side of the uncomplexed carbonyl band is further support for only
oxygen coordination. If there were nitrogen coordination in the complexes, the
nitrogen lone pair would be involved, resulting in a decreased C-N vibration
frequency and a higher-energy carbonyl absorption.
The decrease in the carbonyl stretching frequency of urea upon complexation
to Fe3 +, Cr 3 +, Zn 2 +, or Cu 2 + is interpreted as indicating oxygen coordination
in these complexes.(4" The explanation is similar to that described for the amides.
This conclusion is supported by x-ray studies on the structure of the iron and
chromium complexes. 4 ' Nitrogen coordination is observed in the compounds
Pd(NH 2 CONH 2) 2Cl 2 and Pt(NH2 CONH 2)2C 2, and the spectra show the
expected increase in the C-O stretching frequency as well as a decrease in the
C-N frequency.
Decreases in the P-O stretching frequencies indicative of oxygen coordina-
tion are observed when triphenylphosphine oxide(43 ) and hexamethylphos-
phoramide, OP[N(CH) 2] 31, (4 4 ) are coordinated to metal ions, phenol, or iodine.
A decrease in the S-0 stretching frequency, indicative of oxygen coordination,
is observed when dimethyl sulfoxide or tetramethylene sulfoxide is complexed to
many metal ions, iodine, and phenol.(4 5) The S-O stretching frequency increases
in the palladium complex of dimethyl sulfoxide, compared to free sulfoxide. This
is an indication of sulfur coordination in this complex. The N-0 stretching
frequency of pyridine N-oxide is decreased upon complexation.( 4 6)
The infrared spectra of some ethylenediaminetetraacetic acid complexes
indicate that this ligand behaves as a tetradentate, pentadentate, or hexadentate
ligand in various complexes. The interpretation is based upon absorption bands
in the carbonyl region corresponding to free and complexed carbonyl groups.(47 )
The change in C-N stretching frequency of nitriles(4 8) and metal cyanides(49 )
resulting from their interaction with Lewis acids has attracted considerable
interest. When such compounds are coordinated to Lewis acids that are not able
to 7r back-bond, the C=-N stretching frequency increases. This was originally
thought to be due to the coupling effect described above in the discussion of the
coordination of amides. A combined molecular orbital and normal coordinate
analysis of acetonitrile and some of its adducts( 4 8 ) indicates that a slight increase
is expected from this effect but that the principal contribution to the observed
shift arises from an increase in the CEN force constant. This increase is mainly
due to an increase in the CEN sigma bond strength from nitrogen rehybridiz-
ation.(481 ) In those systems where there is extensive 71 back-bonding from the acid
into the 7t* orbitals of the nitrile group, the decreased pi bond energy accounts
for the decreased frequency.
There have been many papers(5 0 -52) on the assignment of the absorption
bands in simple ammine complexes. The spectra have been analyzed in terms of
The sulfate ion is another good example to demonstrate the effect of change
in symmetry on spectra. The sulfate ion (T, symmetry) has two infrared bands
in the sodium chloride region: one assigned to v3 at 1104 cm-' and one to v4
at 613 cm-'. (See Appendix C for the forms of these vibrations.) In the complex
[Co(NH 3 )sOSO 3 ]Br the coordinated sulfate group has lower symmetry, C 3,- Six
bands now appear at 970 (vi), 438 (v2 ), 1032 to 1044 and 1117 to 1143 (from v3 ),
and 645 and 604 cm- (from v4 ). In a bridged sulfate group the symmetry is
lowered to C 2. and even more bands appear. For a bridged group the v3 band
of SO4 2 - is split into three peaks and the v4 band into three peaks.(6 5) Infrared
spectroscopy is thus a very effective tool for determining the nature of the bonding
of sulfate ion in complexes.
The infrared spectra of various materials containing perchlorate ion have
been interpreted( 66 ) to indicate the existence of coordinated perchlorate. As above,
the change in symmetry brought about by coordination increases the number of
bands in the spectrum. References to similar studies on other anions are contained
in the text by Nakamoto.(18)
In another application of this general idea, it was proposed that the
five-coordinate addition compound, (CH 3 )3 SnCl -(CH 3 )2 SO, had a structure in
which the three methyl groups were in the equatorial positions because the
symmetric Sn-C stretch present in (CH 3)3 SnCl disappeared in the addition
compound.(6 7 ) In (CH 3)3 SnCl the asymmetric and symmetric stretches occur at
545 cm-1 and 514 cm-', respectively.(6 8) In the adduct a single Sn-C vibration
due to the asymmetric Sn-C stretch is detected at 551 cm-'. This is expected
for the isomer with three methyl groups in the equatorial position because a
small dipole moment change is associated with the symmetric stretch in this
isomer. For all other possible structures that can be written, at least two Sn-C
vibrations should be observed. There is also a pronounced decrease in the
frequency of the Sn-Cl stretching mode in the addition compound.
Infrared spectroscopy has been used to excellent advantage in a study of the
influence of pressure on molecular structural transformations. Several systems
were found to change reversibly with pressure. Changes in the infrared spectrum
accompanying high spin-low spin interconversions of transition metal ion systems
with pressure are also reported.(69 )
A very elegant illustration of the consequences of symmetry change on the
infrared spectrum involves the metal carbonyl compounds of general formula
M(CO),X and the 13 C-substituted derivatives.(7 ") Our main interest is the high
wavenumber region associated mainly with the carbonyl stretching vibrations.
This, coupled with the fact that the C=-O stretch and the M-C stretch are
widely separated in energy, enables us to use as a crude basis set for the vibrational
problem the five C-O bond displacement vectors shown in Fig. 6-32A. The
operations of the C,, point group for the all-"C-O compound produce the
result 2A 1 (one radial and one axial), B1 (radial), and E (radial). Three are infrared
active and four are Raman active. The forms of these vibrations are illustrated
in Fig. 6-32B.
The infrared spectrum of Mo(CO),Br is shown in Fig. 6-33. The band labeled
a is assigned to A 1 (radial), e to E1 , and g to A 1 (axial). The assignment of e to
E1 is confirmed by studying(7 " derivatives in which Br is replaced by an
asymmetric group that lowers the symmetry from C,,. This results in splitting
of this band into two peaks separated by 3 to 12 cm -' in various derivatives.
The other bands and shoulders in the spectrum have been assigned to 3 C
6-22 Change in Spectra Accompanying Change in Symmetry upon Coordination 201
o x 0
0 C' C
M
C C
0 C 0
04620
(A)
x x x x x
) X
A1 (radial) A 1 (axial) B1
FIGURE 6-32 (A) CO basis set for carbonyl stretches. (B) The resulting normal
modes.
derivatives, for there is 1.1% "C in naturally occurring carbon compounds.

Consequently, 1.1 % of the molecules have an axial "CO and 4.40% have a radial
"CO. There will be very few molecules with two "CO groups in them.
First, we shall consider the consequences of having an axial "CO in the
molecule. The symmetry is still C,,, and the B, and E modes (which have no
contribution from the axial group) will be unaffected by the mass change. The
two A1 modes having the same symmetry will mix, so isotopic substitution could
affect both; but it will have a major effect on A 1 (axial). The band h (at
1958 cm-1) in the spectrum is assigned to A 1 (axial), for it is roughly what can
be expected for the change in g from the mass change.
(b) (c)
100 a e f g h -100f
80 80-
60 60- - ---
40-4 - -
20 20
0 0---
FIGURE 6-33 (A) High resolution infrared spectrum of Mo(CO) 5 Br in the 2200 to
1900 cm-' region. (B) Spectrum after three hours of exchange with "CO.
X X X X X
A' A A13 A A"

13
FIGURE 6-34 Infrared allowed vibrations in M(CO) 4 ("'CO)X with CO in the radial
position.
For molecules in which a "CO is located in the radial position, the symmetry
is lowered to C, leading to the possible vibrations (the solid circle indicates
"3 CO) illustrated in Fig. 6-34. Only the A" will rigorously be unaffected by the
mass change, and it resembles the E mode in the all-"CO molecule. All of the
other four modes can mix and could be shifted by distributing the mass effect
over all four modes. This would shift the vibrations to lower frequencies than
those in the all "CO molecule. Band f is assigned as an A' mode in the radial
"3 CO monosubstituted molecules that is largely A'. Band d is either a radial A,
mode of the axially substituted molecule or the highest A' of the C, molecule.
The former assignment would involve an intensity that is about I % of the A 1
in the all "CO molecule; since the intensity is found to be about 10%, it is
4
assigned to an A' mode in the C, molecule. A normal coordinate analysis( ) has
been carried out to substantiate these assignments. An important part of this
work is the assignment of a particular band to radial ligands and another to
axial ligands. When "CO is added, the change in intensity of these bands can
be followed and the relative rates of axial versus radial substitution determined.
Jones et al.5 have carried out a normal coordinate analysis on systems of
this type, employing a large number of isotopes. They conclude that the nE-bond
4
orders on these systems obtained by using the Cotton-Kraihanzel( ) approxi-
mations must be viewed with suspicion. The more complete analysis indicates
that as the formal positive charge on the central atom increases in forming
M(CO)sBr from M(CO)6 , the a-bonding increases and the extent of 7 back-
bonding decreases.(5 ) This study(') provides a good reference to illustrate the
information that can be reliably obtained from a normal coordinate analysis.
REFERENCES 1. K. Nakamoto et. al., Chem. Comm., 1451 (1969).

2. E. B.Wilson, J. C. Decius, and P. C. Cross, "Molecular Vibrations," McGraw-Hill, New
CITED York, 1955.
3. G. Barrow, "Introduction to Molecular Spectroscopy," pp. 146-130, McGraw-Hill,
New York, 1962.
4. F. A. Cotton and C. Kraihanzel, J. Amer. Chem. Soc., 84, 4432 (1962); Inorg. Chem.,
2, 533 (1963).
5. L. H. Jones et al., Inorg. Chem., 6, 1269 (1967); ibid., 7, 1681 (1968); ibid., 8, 2349
(1969); ibid., 12, 1051 (1973).
6. a. L. H. Jones and R. R. Ryan, J. Chem. Phys., 52, 2003 (1970).
b. L. H. Jones, "Inorganic Vibrational Spectroscopy," p. 40; M. Dekker, New York,
1971.
c. L. H. Jones and B. I. Swanson, Accts. Chem. Res., 9, 128 (1976).
7. J. Overend and J. R. Scherer, J. Chem. Phys., 32, 1296 (1960).
8. J. 0. Halford, J. Chem. Phys., 24, 830 (1956); S. A. Francis, J. Chem. Phys., 19, 942
(1951); G. Karabatsos, J. Org. Chem., 25, 315 (1960).
9. J. A. Gachter and B. F. Koningstein, J. Opt. Soc. Amer., 63, 892 (1973).
10. J. A. Loader, "Base Laser Raman Spectroscopy," Heyden-Sadtler, London, 1970.
11. a. S. Tobias, "The Raman Effect," Volume 2, Chapter 7, ed. A. Anderson, M. Dekker,
New York, 1973.
b. T. G. Spiro, "Chemical and Biochemical Applications of Lasers," Chapter 2,
C. B. Moore, Ed., Academic Press, New York, 1974.
12. J. Tang and A. C. Albrecht, "Raman Spectroscopy," Chap. 2, ed. H. A. Szymanski,
Plenum Press, New York, 1970.
13. T. G. Spiro and T. C. Strekas, Proc. Nat. Acad. Sci., 69, 2622 (1972); Accts. Chem.
Res., 7, 339 (1974).
14. J. B. Dunn, D. F. Shriver, and I. M. Klotz, Proc. Nat. Acad. Sci., 70, 2582 (1973)
15. National Bureau of Standards, Monograph 70, I to IV.
16. C. H. Townes and A. L. Schawlow, "Microwave Spectroscopy," pp. 56-59, 1110-114,
McGraw-Hill, New York, 1955; Ann. Rev. Phys. Chem.-see the indices for all
volumes.
17. a. W. Gordy and R. L. Cook, "Techniques of Organic Chemistry," Volume IX, Part
II, Microwave Molecular Spectra, ed. A. Weissberger, Interscience, New York, 1970.
b. J. E. Parkin, Ann. Reports Chem. Soc., 64, 181 (1967); ibid., 65, 111 (1968)
c. W. H. Flygare, Ann. Rev. Phys. Chem., 18, 325 (1967).
d. H. D. Rudolph, Ann. Rev. Phys. Chem., 21, 73 (1970).
18. K. Nakamoto, "Infrared Spectra of Inorganic and Coordination Compounds," 2nd
ed., Wiley, New York, 1970. An excellent reference for inorganic compounds.
19. a. M. D. Low, J. Chem. Educ., 47, A 349, A 415 (1970).
b. R. J. Bell, "Introductory Fourier Transform Spectroscopy," Academic Press, New
York, 1972.
c. J. R. Ferraro and L. J. Basils, "FTIR-Applications of Chemical Systems," Volumes
1 and 2, Academic Press, New York, 1978, 1979.
20. M. L. Hair, "Infrared Spectroscopy in Surface Chemistry," M. Dekker, New York, 1967.
21. a. A. T. Bell in "Vibrational Spectroscopies for Adsorbed Species," ed. A. T. Bell and
M. L. Hair, American Chemical Society Symposium Series, 137, 1980.
b. G. Kortum, "Reflectance Spectroscopy: Principles, Methods and Applications,"
s Springer-Verlag, New York, 1969.
c. W. W. Wendlandt and H. G. Hecht, "Reflectance Spectroscopy," Wiley, New York,
1961.
d. W. W. Wendlandt, "Modern Aspects of Reflectance Spectroscopy," Plenum Press,
New York, 1961.
22. a. M. D. Porter, Anal. Chem. 60, 1143A (1988).
b. N. J. Harrick, "Internal Reflection Spectroscopy," Wiley, New York, 1967.
23. a. A. Rosencwaig, "Optoacoustic Spectroscopy and Detection," ed. Y. H. Pao,
Academic Press, New York; Science 181, 657 (1973).
b. D. W. Vidrine, in "Fourier Transform Infrared Spectroscopy," ed. J. R. Ferraro
and I. S. Basile, vol. 3, Chapter 4, Academic Press, New York, 1982.
c. J. W. Chiders and R. W. Palmer, American Laboratory, 22 (1986).
24. a. W. J. Potts, Jr., "Chemical Infrared Spectroscopy," Volume I, Wiley, New York,
1963.
b. R. G. Miller, "Laboratory Methods in Infrared Spectroscopy," Heyden, London,
1965.
25. S. Krimm, "Infrared Spectra of Solids," in "Infrared Spectroscopy and Molecular
Structure," Chapter 8, ed. M. Davies, Elsevier, New York, 1963.
26. a. H. Winston and R. S. Halford, J. Chem. Phys., 17, 607 (1949).
b. D. F. Hornig, J. Chem. Phys., 16, 1063 (1948).
27. M.-L. Josien and N. Fuson, J. Chem. Phys., 22, 1264 (1954).
28. a. A. Allerhand and P. R. Schleyer, J. Amer. Chem. Soc., 85, 371 (1963).
b. R. S. Drago, J. T. Kwon, and R. D. Archer, J. Amer. Chem. Soc., 80, 2667 (1958).
29. a. J. L. Slater et al., J. Chem. Phys., 55, 5129 (1971).
b. H. Huber et al., Nature Phys. Sci., 235, 98 (1972).
c. R. C. Spiker et al., J. Amer. Chem. Soc., 94, 2401 (1972).
d. See G. A. Ozin and A. V. Voet, Accts. Chem. Res., 6, 313 (1973) for a review.
30. a. H. Richert and 0. Glemser, Z. Anorg. Allgem. Chem., 307, 328-344 (1961).
b. N. J. Hawkins, V. W. Coher and W. S. Koski, J. Chem. Phys., 20, 258 (1952).
c. F. X. Powell and E. R. Lippincott, J. Chem. Phys., 32, 1883 (1960).
31. a. M. K. Wilson and G. R. Hunt, Spectrochim. Acta, 16, 570 (1960).
b. E. L. Pace and H. V. Samuelson, J. Chem. Phys., 44, 3682 (1966).
c. R. G. Snyder and J. C. Hisatsune, J. Mol. Spectr., 1, 139 (1957).
32. a. W. C. Price, J. Chem. Phys., 16, 894 (1948).
b. H. H. Claassen, B. Weinstock, and J. G. Maim, J. Chem. Phys., 28, 285 (1958).
c. H. Becher, W. Sawodny, H. N6th, and W. Meister, Z. Anorg. Allgem. Chem., 314,
226 (1962).
33. M. D. Joesten and L. J. Schaad, "Hydrogen Bonding," M. Dekker, New York, 1974
an excellent treatment of the work on hydrogen bonding.
34. M. S. Nozari and R. S. Drago, J. Amer. Chem. Soc., 92, 7086 (1970), and references
therein.
35. R. S. Drago, G. C. Vogel, and T. E. Needham, J. Amer. Chem. Soc.. 93, 6014 (1971);
A. P. Marks and R. S. Drago, J. Amer. Chem. Soc., 97, 3324 (1975).
36. G. C. Vogel and R. S. Drago, J. Amer. Chem. Soc., 92, 5347 (1970).
37. K. F. Purcell and R. S. Drago, J. Amer. Chem. Soc., 89, 2874 (1967).
38. R. S. Drago, B. Wayland, and R. L. Carlson, J. Amer. Chem. Soc., 85, 3125 (1963).
39. F. L. Slejko, R. S. Drago, and D. G. Brown, J. Amer. Chem. Soc., 94, 9210 (1972).
40. C. D. Schmulbach and R. S. Drago, J. Amer. Chem. Soc., 82, 4484 (1960); R. S. Drago
and D. A. Wenz, J. Amer. Chem. Soc., 84, 526 (1962).
41. R. B. Penland, S. Mizushima, C. Curran, and J. V. Quagliano, J. Amer. Chem. Soc.,
79, 1575 (1957).
42. Y. Okaya et al., "Abstracts of Papers of 4th International Congress of the International
Union of Crystallography," p. 69, Montreal, 1957.
43. F. A. Cotton et al., J. Chem. Soc., 1961, 2298, 3735, and references contained therein.
44. J. T. Donoghue and R. S. Drago, Inorg. Chem., 1, 866 (1962).
45. R. S. Drago and D. W. Meeks, J. Phys. Chem., 65, 1446 (1961), and papers cited therein.
46. J. V. Quagliano et al., J. Amer. Chem. Soc., 83, 3770 (1961).
47. M. L. Morris and D. H. Busch, J. Amer. Chem. Soc., 78, 5178 (1956).
48. K. F. Purcell and R. S. Drago, J. Amer. Chem. Soc., 88, 919 (1966).
49. D. F. Shriver and J. Posner, J. Amer. Chem. Soc., 88, 1672 (1966).
50. M. G. Miles et al., Inorg. Chem., 7, 1721 (1968), and references therein.
51. S. Mizushima, I. Nakagawa, and J. V. Quagliano, J. Chem. Phys., 23, 1367 (1955); G.
Barrow, R. Krueger, and F. Basolo, J. Inorg. Nucl. Chem., 2, 340 (1956).
52. J. Chatt, L. A. Duncanson, and L. M. Venanzi, J. Chem. Soc., 1955, 4461; 1956; 2712.
53. J. Lewis, R. S. Nyholm, and P. W. Smith, J. Chem. Soc., 1961, 4590.
54. A. Sabatini and I. Bertini, Inorg. Chem., 4, 959 (1965).
55. N. J. DeStefano and J., L. Burmeister, Inorg. Chem., 10, 998 (1971).
56. J. L. Burmeister, Coord. Chem. Rev., 3, 225 (1968).
57. A. H. Norbury and A. I. P. Sinha, Quart. Rev. Chem. Soc., 24, 69 (1970).
58. A. D. Allen and C. V. Senoff, Chem. Comm., 621 (1965); A. D. Allen, et al., J. Amer.
Chem. Soc., 89, 5595 (1967); A. D. Allen and F. Bottomley, Accts. Chem. Res., 1, 360
(1968).
59. J. P. Collman and Y. Kang, J. Amer. Chem. Soc., 88, 3459 (1966).
60. I. Yamamoto et al., Chem. Comm., 79 (1967); Bull. Chem. Soc. Japan, 40, 700 (1967).
61. P. B. Chock and J. Halpern. J. Amer. Chem. Soc., 88, 3511 (1966).
62. L. Vaska, Chem. Comm., 614 (1960).
63. B. M. Gatehouse, S. E. Livingston, and R. S. Nyholm, J. Chem. Soc., 1957, 4222; C.
C. Addison and B. M. Gatehouse, J. Chem. Soc., 1960, 613.
64. J. R. Ferraro, J. Inorg. Nucl. Chem., 10, 319 (1959).
65. K. Nakamoto et al., J. Amer. Chem. Soc., 79, 4904 (1957); C. G. Barraclough and M.
L. Tobe, J. Chem. Soc., 1961, 1993.
66. B. J. Hathaway and A. E. Underhill, J. Chem. Soc., 1961, 3091.
67. N. A. Matwiyoff and R. S. Drago, Inorg. Chem., 3, 337 (1964).
68. H. Kriegsman and S. Pischtschan, Z. Anorg. Allgem. Chem., 308, 212 (1961); W. F.
Edgell and C. H. Ward, J. Mol. Spectr., 8, 343 (1962).
69. J. R. Ferraro and G. J. Long, Accts. Chem. Res., 8, 171 (1975).
70. H. D. Kaesz et al., J. Amer. Chem. Soc., 89, 2844 (1967).
71. J. B. Wilford and F. G. A. Stone, Inorg. Chem., 4, 389 (1965).
A. In the area of infrared, highly recommended references dealing with theory are:
ADDITIONAL
REFERENCES
Introductory
N. B. Colthup, L. H. Daly, and S. E. Wiberley, "Introduction to Infrared and Raman

Spectroscopy," 3rd ed., Academic Press, New York, 1990.
L. H. Jones, "Inorganic Vibrational Spectroscopy," Vol. 1, M. Dekker, New York, 1971.
W. G. Fately et al., "Infrared and Raman Selection Rules for Molecular and Lattice
Vibrations: The Correlation Method," Wiley-Interscience, New York, 1972.
L. A. Woodward, "Introduction to the Theory of Molecular Vibrations and Vibrational
Spectroscopy," Oxford University Press, New York, 1972.
G. Barrow, "Introduction to Molecular Spectroscopy," McGraw-Hill, New York, 1962.
Advanced
G. Herzberg, "Spectra of Diatomic Molecules," 2nd ed., Van Nostrand, Princeton, 1950.
G. Herzberg, "Infrared and Raman Spectra of Polyatomic Molecules," Van Nostrand,
Princeton, 1945.
E. B. Wilson, J. C. Decius, and P. C. Cross, "Molecular Vibrations," McGraw-Hill, New
York, 1955.
D. Steele, "Theory of Vibrational Spectroscopy," W. B. Saunders, Philadelphia, 1971.
B. References dealing with applications are:
L. H. Jones, "Inorganic Vibrational Spectroscopy," Vol. 1, M. Dekker, New York, 1971.

M. Davies, ed., "Infrared Spectroscopy and Molecular Structure," Elsevier, Amsterdam,
1963.
K. Nakamoto, "Infrared and Raman Spectra of Inorganic and Coordination Compounds,"
4th ed., Wiley, New York, 1986.
K. Nakanishi and P. H. Solomon, "Infrared Absorption Spectroscopy" 2nd ed., Holden-
Day, San Francisco, 1977. This reference contains many examples of involving
structure determination of organic compounds from infrared spectra.
A. Finch et al., "Chemical Applications of Far Infrared Spectroscopy," Academic Press,
New York, 1970.
L. J. Bellamy, "The Infrared Spectra of Complex Molecules," 3rd ed., Wiley, New York,
1975.
L. J Bellamy, "Advances in Infrared Group Frequencies," Halsted Press, New York, 1968.
J. R. Ferraro, "Low Frequency Vibrations of Inorganic Coordination Compounds,"

"Practical Fourier Transform Infrared Spectroscopy," eds. J. R. Ferraro and K. Krishman,
Academic Press, San Diego, 1990.
Compilations
"Sadtler Standard Spectra," Sadtler Research Laboratories, Philadelphia, Pa. (A collection

of spectra for use in identification, determination of the purity of compounds, etc.)
Raman Spectroscopy
H. A. Szymanski, ed., "Raman Spectroscopy," Plenum Press, New York, 1967; Vol. 2,1970.
A. Anderson, ed., "The Raman Effect," Vols. 1 and 2, M. Dekker, New York, 1971.
M. C. Tobin, "Laser Raman Spectroscopy," Wiley, New York, 1971; Vol. 35 in "Chemical
Analysis," P. J. Elving and I. M. Kolthoff, ed.
R. S. Tobias, J. Chem. Educ., 44, 2, 70 (1967).
J.A. Koningstein, "Introduction to the Theory of the Raman Effect," D. Reidel, Dordrecht,
1972.
D. A. Long, "Raman Spectroscopy," McGraw-Hill, New York, 1977.
J. M. Stencel, "Raman Spectroscopy for Catalysis," Van Nostrand Reinhold, New York,
1990.
A. T. Tu, "Raman Spectroscopy in Biology" Wiley, New York, 1982.
C. B. Moore, ed., "Chemical and Biochemical Applications of Lasers," Academic Press,
New York, 1974.
EXERCISES 1.Consider the C3 , structure of N-SF 3. (a) Show that the total representation is E = 15;
2C 3 = 0; 3u, = 3. (b) Indicate the procedure that shows that the irreducible represen-
tations 4A 1 , 1A2 and 5E result from this total representation. (c) Show that
4Ai + 1A2 + 5E equals E = 15; 2C 3 = 0; 30 = 3. (d) Indicate the species for the
allowed: (e) infrared, (f) Raman, and (g) Raman-polarized bands. (h) How many bands
would be seen for (e), (f), and (g), assuming all allowed fundamentals are observed?
2. Consider XeF 4 as having D4, symmetry. Using the procedure in exercise 1, show that
the total number of vibrations is indicated by A g + B,, + Byg + A2 . + B 2 u+ 2E.
Also indicate the species for the allowed (a) infrared, (b) Raman, and (c) Raman-
polarized bands. (d) How many bands would be seen for (a), (b), and (c), assuming all
allowed fundamentals are observed?
3. Consider the molecule trans-N 2F 2.
a. To which point group does it belong?
b. How many fundamental vibrations are expected?
c. To what irreducible representations do these belong?
d. What is the difference between the A, and B. vibrations?
e. Which vibrations are infrared active and which are Raman active?
f. How many polarized lines are expected?

Exercises 207
g. How many lines are coincident in the infrared and Raman? Does this agree with
the center of symmetry rule?
h. To which species do the following vibrations belong? N-N stretch, symmetric

N-F stretch, and asymmetric N-F stretch.
i. Can the N-N stretch and N-F stretch couple?
4. Indicate to which irreducible representations the vibrations belong and indicate how
many active IR, active Raman, and polarized Raman lines are expected for:
a. cis-N 2 F2 (C2 ,)
b. A linear N 2 F 2 molecule (D.h)
5. a. Why is cis-N 2 F 2 in the point group C , instead of D h?

2 2
b. What symmetry element is missing from the pyramidal structure of SF 4 that is

required for D 4 . (i.e., which one element would give rise to all missing operations)?
6. The infrared spectrum of gaseous HCl consists of a series of lines spaced 20.68 cm-
apart. (Recall that wavenumbers must be converted to frequency to employ the
equations given in this text.)
a. Calculate the moment of inertia of HCl.
b. Calculate the equilibrium internuclear separation.
7. Suppose that valence considerations enabled one to conclude that the following
structures were possible for the hypothetical molecule X2Y 2 :
Y Y Y Y
Y--X-X--Y X-X X-X X--X
Y Y -
The infrared spectrum of the gas has several bands with P-, Q-, and R-branches. Which
structures are eliminated?
8. An X-H fundamental vibration in the linear molecule A-X-H is found to occur

at 3025 cm-'. At what frequency is the X-D vibration expected? It is found that
the X-D vibration is lowered by only about half the expected amount, and the A-X
stretching frequency is affected by deuteration. Explain. Would you expect the
A-X-H bending frequency to be affected by deuteration? Why?
9. How many normal modes of vibration does a planar BC]3 molecule have? Refer to
Appendix C and illustrate these modes. Indicate which are infrared inactive. Confirm
these conclusions by employing symmetry considerations to determine the number of
active infrared and raman lines. Indicate the Raman polarized and depolarized lines
and the parallel and perpendicular vibrations.
10. The following assignments are reported for the spectrum of GeH 4 : 2114, T2 ; 2106,
A,; 931, E; 819, T2. Label these using the v, symbolism. (E modes are numbered after
singly degenerate symmetric and asymmetric vibrations, and T modes are numbered
after E modes.) Refer to Appendix C to illustrate the form for each of these vibrations
and label them as bends or stretches.
11. Refer to the normal modes for a planar ZXY 2 -molecule (Appendix C). Verify the
assignment of v, as B2 and v, as B1 by use of the character tables. Explain the procedure.
12. a. Indicate the infrared and Raman activities for the modes of the trigonal bipyramidal
XY 5 -molecule (Appendix C).
b Does this molecule possess a center of symmetry?
13. The spectrum of Co(NH 3 )(Cl0 4 )3 has absorption bands at 3320, 3240, 1630, 1352,
and 803 cm-'. For purposes of assignment of the ammonia vibrations, the molecule
can be treated as a C 3 ,-molecule, Co-N-H Refer to Appendix C for a diagram
of these modes; utilizing the material in this chapter and Chapter 4, assign these
modes. Use the v. symbolism to label the bands and also describe them as bends,
stretches, etc.
14. What changes would you expect to see in the infrared spectrum of CH COSCH if
3 3
a. coordination occurred on oxygen?
b. coordination occurred on sulfur?
15. In which case would the spectrum of coordinated [(CH3) 2 N] 3 PO be most likely to
resemble the spectrum of the free ligand:
a. Oxygen coordination?
b. Nitrogen coordination?
Why?
16. Using a value of k = 4.5 x 10' dynes cm 'for the C-C bond stretching force constant,
calculate the wavenumber (cm-1) for the C-C stretching vibration.
CH 2 -CH 2
17. The compound HgCl 2 -O CH2--CH O could be a linear polymer with dioxane
in the chair form or a monomer with bidentate dioxane in the boat form. How could
infrared and Raman be used to distinguish between these possibilities?
18. Which complex would have more N-O vibrations: (NH 3 ),CoNO 2 2 + or
2
(NH 3 )sCoONO +?
19. What differences (number of bands and frequencies) are expected between the C-O
absorptions of the following:
0
CH 3 COO-, Zn C-CH -2H20, CH 3 COOH, and CH 3 C 7 Ag?
20. Given the NH 3 molecule below, determine the total representation for the vibrations
using r's and O's (02 not shown). Use projection operators to determine the vibrational
wave functions.
Exercises 209
21. Consider the tetrahedral anion BCl4
a. How many bands do you expect in the infrared spectrum?
b. How many bands do you expect in the Raman spectrum?
c. How many Raman-polarized bands do you expect?
d. How many B-Cl stretches do you expect in the IR?
22. The hydrogen bonding Lewis acid CHCI3 has a C-H stretching frequency that is
lowered upon deuteration.
a. What is responsible for the theoretical factor of 1.4?
b. What could be responsible if the observed factor were less than 1.4?
c. When the compound is dissolved in the Lewis base NEt3 , the C-H stretching
frequency is lowered. Rationalize the lowering.
d. If one examines the C-D shift upon adduct formation with EtaN, the percentage
lowering is smaller than that in (b). Rationalize this finding.
23. The compound (CH 3 )3 SnCl exhibits two infrared active Sn-C stretches at 545 cm
and 514 cm -. The IR spectrum on the five-coordinate adduct B -(CH 3 )3 SnCl (where
B is a base) shows a single Sn-C stretching band around 550 cm-'.
a. Consider a tetrahedral structure for (CH 3 )3 SnCl and predict the number of Sn-C
stretches expected.
b. Assume a trigonal bipyramidal structure for the adduct with the methyl groups
lying in a plane. How many Sn-C stretching bands are expected?
c. How do you explain the actual spectrum?
24. a. How many infrared-active normal modes containing Sn-Cl stretching would be
expected for the six-coordinate species cis-SnCI4 X 2 ?
b. Describe briefly the selection rules for Raman spectrosocopy and for infrared
spectroscopy.
25. Iron pentacarbonyl, Fe(CO) 5 , possesses D 3 h symmetry.
a. Determine the number and irreducible representations of the IR-active and Raman-
active fundamentals to be expected for this compound.
b. Determine the number of IR-active carbonyl stretching bands to be expected.
26. Consider the following IR data, taken from Nakamoto's book" 8 ):
K 3 [Mn(CN) 6 ] 2125 cm-1

K 4 [Mn(CN)6] 2060 cm-'
K,[Mn(CN),] 2048 cm-'
Discuss this trend in CN stretching frequencies in terms of bonding in the complexes,

and the relationship this has on force constants and frequencies. What do you predict
for the Mn-C force constants in this series?
27. Use of IR and Raman spectroscopy is valuable in determining the stereochemistry of

metal carbonyl complexes and various substituted metal carbonyls. C=O stretches
are often analyzed separately from the rest of the molecule (explain why this assumption
is or is not valid).
a. For Cr(CO), and Cr(CO),L (treat L as a point ligand), work out the symmetry
of the IR- and Raman-allowed CO stretches.
b. Do the same as in (a) for Mo(CO)4 DTH (DTH is the bidentate ligand
CH 3 SCH 2 CH2 SCH 3 ), and for trans-Mo(CO)4 [P(OCH,)3 ]2 . Do your findings
explain the IR spectra below? (Hint: remember asumptions implicit in the
analysis.) [Spectra reported by M. Y. Darensbourg and D. J. Darensbourg, J.
Chem. Ed., 47, 33 (1970).]
2000 1900 cm-1 2000 1900 cm 1

Lu LU
0 0
Z- 80 Z -80
- 60 - 60
zU) Z 0)
S40 4-40
I- I-
Nuclear Magnetic
Resonance
Spectroscopy-
7
Elementary Aspects
In this chapter, the principles necessary for an elementary appreciation of nuclear Introduction(1 ,2,3)
magnetic resonance' (3) (nmr) will be presented. The reader mastering this
chapter will have a minimum knowledge of those principles required for spectral
interpretation of the results from slow passage as well as Fourier transform
experiments. Chapter 8 expands upon several of the important concepts in-
troduced here and illustrates several other types of applications of nmr. Varying
appreciations of this subject can be obtained by complete reading of Chapter 7
and selected readings from Chapter 8.
Protons and neutrons both have a spin quantum number of /2 and,
depending on how these particles pair up in the nucleus, the resultant nucleus
may or may not have a net non-zero nuclear spin quantum number, I. If the
spins of all the particles are paired, there will be no net spin and the nuclear spin
quantum number I will be zero. This type of nucleus is said to have zero spin
and is represented in Fig. 7M1(A). When I is '/2, there is one net unpaired spin and
this unpaired spin imparts a nuclear magnetic moment, p, to the nucleus. The
distribution of positive charge in a nucleus of thtitype is spherical. The properties
for I = /2 are represented symbolically in Fig. 7-1(B) as a spinning sphere (vide
infra). When I > 1, the nucleus has spin associated with it and the nuclear charge
distribution is non-spherical; see Fig. 7-1(C). The nucleus is said to possess an
electric quadrupole moment eQ, where e is the unit of electrostatic charge and
/=0, P= 0 /=!, p 0, eQ= 0 />10,*0 eQ>0 />1,p*0, eO<O
(A) (B) (C) (D)
FIGURE 7-1 Various representations of nuclei.

212 Chapter 7 Nuclear Magnetic Resonance Spectroscopy-Elementary Aspects
Q is a measure of the deviation of the nuclear charge distribution from spherical

symmetry. For a spherical nucleus, eQ is zero. A positive value of Q indicates
that charge is oriented along the direction of the principal axis (Fig. 7-1(C)),
while a negative value of Q indicates charge accumulation perpendicular to the
principal axis (Fig. 7-1(D)).
Nuclei with even numbers of both protons and neutrons all belong to the
type rereisinTed in-Fig. 7-1(A). Typical examples include 160, 12C, and SThe
values for I, p, and eQ for many other nuclei have been tabulated,(, 2 ,3 ) and are
more easily looked up than figured out. A table summarizing these properties is
included inside the front cover.
Nuclear magnetic resonance spectroscopy is often concerned with nuclei
with I = '/2, examples of which include 'H, 3 C, 3 P, and 9 F. Spectra also result
from nuclei for which I ; 1, but cannot be obtained on nuclei with I = 0.
Unpaired nuclear spin leads to a nuclear magnetic moment. The allowed
orientations of the nuclear magnetic moment vector in a magnetic field are
indicated by the nuclear spin angular momentum quantum number, n. This
quantum number takes on values I, I - 1,...,(-I+ 1), -I. When I2,
MI = ±'/2 corresponding to alignments of the magnetic moment with and op-
posed to the field. When I = 1, m, has values of 1, 0, and -1, corresponding,
respectively, to alignments with, perpendicular to, and opposed to the field. In
the absence of a magnetic field, all orientations of the nuclear moment are
degenerate. In the presence of an external field, however, this degeneracy will be
removed. For a nucleus with I = 1/2, the mr = +1/2 state will be lower in energy
and the -'/2 state higher, as indicated in Fig. 7-2. The energy difference between
the two states, at magnetic field strengths commonly employed, corresponds to
radio frequency radiation; it is this transition that occurs in the nmr experiment.
1,+.M = 1
22
AE
No applied Applied
field magnetic field
FIGURE 7-2 Splitting of the n = + 2 states in a magnetic
field.
Classical 7-1 SOME DEFINITIONS

Description of We begin our classical description of nmr by reviewing some of the background
the NMR physics of magnetism and defining a few terms necessary to understand nmr. It
is very important to appreciate what is meant by angular momentum. Circulating
Experiment-The charges have angular momentum associated with them. Planar angular mo-
Bloch Equations mentum, p(cp) is given by (p) = Y x mv where-f, shown in Fig. 7-3, is the position
7-2 Behavior of a Bar Magnet in a Magnetic Field 213
vector of the particle e; v its linear momentum vector; m is the mass; and p is
the angular change that signifies an angular momentum. The angular momentum r,
p is perpendicular to the plane of the circulating change, and the direction of
the angular momentum vector is given by the right-hand rule. For the situation
shown in Fig. 7-3, the vector points into the page.
The nucleus is a more complicated three-dimensional problem. The total FIGURE 7-3 Circulating
angular momentum of a nucleus is given by J, but it is convenient to define a charge giving rise to planar
dimensionless angular momentum operator I by angular momentum.
i=hi or f1 = /h (7-1)
Associated with the angular momentum is a classical magnetic moment, AN,

which can be taken as parallel to J, so:
AN = J = 1hl (7-2)
where T, the magnetogyric (sometimes called gyromagnetic) ratio is a constant

characteristic of the nucleus. From equation (7-2), we see that the magnetogyric
ratio represents the ratio of the nuclear magnetic moment to the nuclear angular
momentumn.
7-2 BEHAVIOR OF A BAR MAGNET

IN A MAGNETIC FIELD
A nucleus with a magnetic moment can be treated as though it were a bar magnet.
If a bar magnet were placed in a magnetic field, the magnet would precess about
the applied field, Ho, as shown for a spinning nuclear moment in Fig. 7-4. Here
o is the angle that the magnetic moment vector makes with the applied field,
and w, the Larmor frequency, is the frequency of the nuclear moment precession.
The instantaneous motion of the nuclear moment (indicated by the arrowhead
on the dashed circle) is tangential to the circle and perpendicular to A and H0 .
The magnetic field is exerting a force or torque on the nuclear moment, causing
it to precess about the applied field. For use later, we would like to write an
equation to describe the precession of a magnet in a magnetic field. The applied
field fl exerts a torque - on A which is given by the cross product:
T- = x No (7-3)
moment
The right-hand rule tells you that the moment is precessing clockwise in Fig. 7-4,
with the torque and instantaneousmotion perpendicular to A and H0o. According
to Newton's law, force is equal to the time derivative of the momentum, so the
torque is the time derivative of the angular momentum.
FIGURE 7-4 Precession of
d( a nuclear moment in an
dt applied field of strength Ho.
With the equation yhI,

d the equationforprecession of the moment is given
by:
dii yd(hI)
dt dt
or (7-4)
d-t
- p#= -yHO xp
dt
(The dot is an abbreviation for the time derivative of some property, in this case
#; the minus sign arises because we have changed the order for taking the
cross product.) Thus, o, the precession frequency or the Larmor frequency, is
given by:
o = Ho (7-5)
where the sign of 7 determines the sense of the precession. According to equation
(7-5), the frequency of the precession depends upon the appliedfield strength and
the magnetogyric ratio of the nucleus. The energy of this system is given by the dot
product of i and H.
E - Ho -p||Ho0 cos 0 (7-6)
7-3 ROTATING AXIS SYSTEMS
There is one more mathematical construct that greatly aids the analysis of the
nmr experiment, and this is the idea of a rotating coordinate system or rotating
frame. In Fig. 75, a set of x-, y-, and z-coordinates is illustrated. The rotating
frame is described by the rotating axes u and v, which rotate at some frequency
z uw, in the xy-plane. The z-axis is common to both coordinate systems. If the
rotating frame rotates at some frequency less than the Larmor frequency, it would
appear to an observer on the uv frame that the precessional frequency has slowed,
v which would correspond to a weakening of the applied field. Labeling this
apparently weaker field as He,, our precessing moment is described in the
rotating frame by
y
=t -yHX (7-7)
where |Hef I <I H,. Heff is seen to be a function of w, and if w is faster than
U the precessional frequency of the moment, w, it will appear as though the z-field
has changed direction. The effective field, Heff, can be written as:
frame u, v, z in a CartesianI1
coordinate system x, ff Ho - (7-8)
y, z.
7-4 Magnetization Vectors and Relaxation 215
where 22 is a unit vector along the z-axis. The time dependence of the moment,
#, can be rewritten as
it = -- y H, - '0)x] x (7-9)
When w = w = 7H 0 , then |H.,|I= 0 and # appears to be stationary. Thus, in-

a frame rotating with the Larmor frequency, we have eliminated time dependency
and, in so doing, greatly simplified the problem.
7-4 MAGNETIZATION VECTORS AND RELAXATION
The concepts in this section are vital to an understanding of nmr. Of these, the
concept of relaxation will be developed in stages as the necessary background
information is covered. For practical purposes, it is necessary to consider an
ensemble or large number of moments because the nmr experiment is done with
bulk samples. The individual moments in the sample add vectorially to give a
net magnetization, M.
NiZii, (7-10)
In an ensemble of spins in a field, those orientations aligned with the field will
be lower in energy and preferred. However, thermal energies oppose total
alignment; experimentally, only a small net magnetization is observed. The
equation for the motion of precession of M is similar to that for #, i.e.,
M =- yH x M (7-11)
If we place a sample in a magnetic field at constant temperature and allow

the system to come to equilibrium, the resulting system is said to be at thermal
equilibrium. At thermal equilibrium, the magnetization, M,, is given by
Mol = Nk I(I + 1)HO (7-12)

3k T
where k is the Boltzmann constant and No is the number of magnetic moments

per unit volume (i.e., the number of spins per unit volume). This equation results
from equation (7-6) and the assumption of a Boltzmann distribution. For H 2 0
at 300 K and Ho = 10,000 gauss (1 tesla) the value of M, is about 3 x 10-6 gauss A
(3 x 10- "T).
The description of the magnetization is not yet complete for, at equilibrium, \1
we have a dynamic situation in which any one given nuclear moment is rapidly
changing its orientation with respect to the field. The mechanism for reorientation
involves time-dependent fields that arise from the molecular motion of magnetic
nuclei in the sample. Suppose, as in Fig. 7-6, that nucleus B, a nucleus that has
B
a magnetic moment, passed by nucleus A, whose nuclear moment is opposed to
the field. The spin of A could be oriented with the field via the fluctuating field FIGURE 7-6 Relaxation
mechanisms.
from moving B. In the process, the translational or rotational energy of the
(A) (B) (C) (D) (E) (F)
FIGURE 7-7 The decay of the M,-component and growth of the

Mzcomponent when a field in the u-direction is turned off and one in the
z-direction is turned on.
molecule containing B would be increased, since A gives up energy on going to

a more stable orientation aligned with the field. This process is referred to as
spin-lattice relaxation. Nucleus B is the lattice; it can be a magnetic nucleus in
the same or another sample molecule or in the solvent. These nuclei do not have
to be hydrogens and may even be unpaired electrons in the same or other
molecules. The fluctuating field from motion of the magnetic nucleus must have
the same frequency as that of the nmr transition in order for this relaxation
process to occur. However, moving magnetic nuclei have a wide distribution of
frequency components, and the required one is always present.
Another process can also occur when the two nuclei interact, whereby
nucleus B, with m, = 1/2 goes to the higher-energy m- 1/2 state while nucleus
A changes from - /2 to + 1/2 There is no net change in spin from this process,
and it is referred to as a spin-spin relaxation mechanism.
To gain more insight into the nature of relaxation processes, let us examine
the decay in magnetization as a function of time. Consider an experiment in
which we have our magnetization aligned along the u axis as in Fig. 7-7(A). We
then switch on a field aligned with the z-axis; Mu diminishes and M. grows,
generally at different rates. In Fig. 7-7(E), for example, the u-component has
completely decayed, but the z-component has not yet reached equilibrium, for
it gets larger in Fig. 7-7(F). The growth and decay are first-order processes, and
Bloch proposed the following equation for the three-dimensional case (i.e., u-, v-,
and z-magnetizations are involved and the field is located along the z-axis):
-M - M
T2 1T2
and M - - (MZ - MO) (7-14)

T
where 1/T2 and 1/T are first-order rate constants, Mo is the equilibrium value
of the z-magnetization, and u- and v-components vanish at equilibrium. In
general, it is found that 1/T2 ; 1/T1 . Instead of rate constants, 1/T, it is more
usual to refer to their reciprocals (i.e., the T's), which are lifetimes or relaxation
times. Since Ti refers to the z-component, it is called the longitudinal relaxation
time, while T2 is called the transverse relaxation time. The spin-lattice mechanism
contributes to T, and the spin-spin mechanism is one of several contributions
to T2.
7-5 The NMR Transition 217
7-5 THE NMR TRANSITION
Before proceeding with our classical description of the nmr experiment, it is

advantageous to introduce a few concepts from the quantum mechanical descrip-
tion of the experiment. When the bare nucleus (no electrons around it) is placed
in a magnetic field, H,, the field and the nuclear moment interact [see equation
(7-6)] as described by the Hamiltonian for the system
H= ---jN o (7-15)
where, as shown in equation (7-2), IN = yhI with the N subscript denoting a

nuclear moment. When it is obvious that we are referring to a nuclear moment,
the N subscript will be dropped. Combining equations (7-2)* and (7-15) yields
A - - = g Ho
NNhHo (7-16)
where yN is constant for a given nucleus, gN is the nuclear g factor and:
f#N = eb 2mc (7-17)
In equation (7-17), m is the mass of the proton, e is the charge of the proton,
and c is the speed of light.
The expectation values of the operator f., where z is selected as the applied
field direction, are given by m, where m, = I, I 1. T.-he degeneracy of-,
+1
the m1 states that existed in the absence of the field is removed by the interaction +2
between the field H, and the nuclear magnetic moment pN. The quantized
orientations of these nuclear moments relative to an applied field H0 are shown
in Fig. 7-8 for I = /2 and I = 1. Since the eigenvalues of the operator I are m,
the eigenvalues of R (i.e., the energy levels) are given by equation (7-18).
1
2
E =- yhmHo (7-18)
(A)
The energy as a function of the field is illustrated for I = /2 in Fig. 7-9. The
quantity gN is positive for a proton, as is N, so the positive m, value is lowest
in energy. The nuclear wave functions are abbreviated as lY> and |#> for the +'/2/ +1
and - /2 states, respectively. By inserting values of m, = + /2 and m, = - /2
into equation (7-18), we find that the energy difference, AE, between these states 0
3
or the energy of the transition hv for a nucleus of spin '/2 is given by gN NHo
or yhHo.
--
AE = gN#NHO (7-18a)
(B)
FIGURE 7-8 Quantized

* We locate the field Ho along the z-axis and, since H, and H, are zero, only Hof, is non-zero; orientation of m, for (A)
i.e., HjI. and HY, are zero because H. and H, are zero. 1 ='2 and (B) I = 1.
M1- 1
E 2
N
FIGURE 7-9 The field dependence of E hi = = y11H,
R -- g=9/3NHO= AH
the energies of m, = t1/2. G
Y
m, = + 2
0 -H inc.->
Since m, = + 1/2 is lower in energy than m1 = -/2, there will be a slight

excess population of the low energy state at room temperature, as described by
the Boltzmann distribution expression in equation (7-19):
N(-/2) AE
= exp(- AE/kt) ~ 1 -- - when AE < kT (7-19)
N(+'/2) kT
For a proton, AE is ~-10' cm- in a 1 T field, and kT - 200 cm-'. At room

temperature, there is a ratio of 1.0000066 (+'/2) spins to one (_'/2). The
probability of a nuclear moment being in the +'/2 state is (/2) 1+ MHO and
kT)
that of it being in the - '/2 state is (/2) 1 + y2) (recall our discussion of
kT)
equation (7-12)).
The energy separation corresponding to hv occurs in the radio-frequency
region of the spectrum at the magnetic field strengths usually employed in the
experiment. One applies a circularly polarized radio frequency (r.f.) field at right
angles to Ho (vide inf-a), and the magnetic component of this electromagnetic
field, H, provides a torque to flip the moments from m1 = +/2 to -2', causing
transitions to occur.
7-6 THE BLOCH EQUATIONS
In order to understand many of the applications of nmr, it is necessary to

appreciate the change in magnetization of the system with time as the H, field
is applied. This result is provided by the Bloch equation. Incorporating equations
(7-13) and (7-14), describing relaxation processes, into equation (7-11), which
describes the precession of the magnetization, and converting to the rotating
frame gives the Bloch equation:
-1 - - 1 1
M = -7Hef x M - - (M.e. + M j -- (M. - Mo)e, (7-20)
T2 T1
torque from relaxation effects

the magnetic
field
7-6 The Bloch Equations 219
In the presence of H, and in the rotating frame, equation (7-8)-which describes
Hef in the rotating frame-becomes:
Hef = (HO - 2 + H1 , (7-21)
where the fra e is rotating with the frequency w corresponding to the frequency
of the H 1 field, the oscillating field at right angles to Ho. Equation (7-20) is a
vector equation in the rotating frame that can best be written in terms of the
components of M, which are M., M,, and M2.
The three components of the Bloch equation are:
. dM
dt (w0 - w )M, " (7-22)
T2
. dM,
w'i)M + 1yH 1M2 (7-23)
dt T2
. dM.
M2 - -d -- H1 M, + (Mo - M)/T (7-24)
dt
In these equations, wo is the Larmor frequency, which equals IH 0, and the

u, v reference frame is rotating at angular velocity wi.
Experimentally, we monitor the magnetization in the xy-plane, referred to
as the transverse component. Using a phase-sensitive detector, we monitor the
component of magnetization induced along the u-axis. In a slow-passage or
steady-state experiment, only a u-component of magnetization exists; but because
of a 90 phase lag associated with the electronic detection system, a component
90 out of phase with u is measured. Slow-passage or steady-state conditions
require H, to be weak (of the order of milligauss) compared to H, (which is of
the order of kilogauss). Then, according to equation (7-21), the z-component
dominates unless wo is close to w so that the first term (H = w=/7) becomes
Net
small. (Note that wi is the frequency of the r.f. field and not the Larmor frequency magnetization
of precession about H1 ; i.e., wi # y/H1 .) When wi is close to the Larmor
frequency, wo, then Hef, is tipped toward the u-axis. Since H, is being changed
slowly, the net effect is to change Hef, slowly. The individual moments continue
to precess about Hff as a consequence of the torque, which is perpendicular to
H,,,. As a result, M remains parallel to Hfl, and a u-component results. Figure
(M, is the net z component
7-10 represents the tipping of M to remain aligned with He, as we pass through of the magnetization and
resonance. The H1 field is along the u-axis. It is also an alternating field with not a precessing magnetic
the frequency of the rotating frame. A static field will not tip the magnetization moment)
vector significantly because H 1 is so small. FIGURE 7-10 Effect of H, on
Relaxation processes, T, try to preserve the orientation along the strong the net magnetization of an
z-field, as shown by the arrow labeled T in Fig. 7-10. Under steady-state HO field.
conditions, all time derivatives are zero, so equations (7-22), (7-23), and (7-24)
are all equal to zero and can be solved to produce:
M. = MO TH2 (wo - w ) 1 (7-25)

1+ T - w1)2 + y 2 H 1 TT
((w 2
MH = M 7HI T2 2 (7-26)
+ T2 (O -o_) + 2
H 12 T1 T2
1 + T 2 (w0 - w)2
M, = Mo 2(O 2 (7-27)
1 + T/(w0 - w,)2 + H1 2 T1 T2
7-7 THE NMR EXPERIMENT
Equations (7-25) to (7-27) describe the magnetization of our sample in the

so-called slow passage experiment, which is schematically illustrated in Fig. 7-11.
In this method, one applies a strong homogeneous magnetic field, causing the
nuclei to precess. Radiation of energy comparable to E is then imposed with
a radio frequency transmitter, producing H 1 . Whe he applied frequency from
the radio transmitter is equal to the Larmor f quency, the two are said to be
in resonance, and a u,v-component is induced which can be detected. This is
the condition in (7-21) when HO wi/y. Quantum mechanically, this is equiv-
alent to some nuclei being excited from the low energy state (m = + 1/2) to the
high energy state (m, = -1/2; see Fig. 7-9) by absorption of energy from the r.f.
source at a frequency equal to the Larmor frequency. Since AE = hv and 9 = 27v,
AE is proportional to the Larmor frequency, w. Energy will be extracted from
the r.f. source only when this resonance condition (a = 27[v) is fulfilled. With an
electronic detector (see Fig. 7-11), one can observe the frequency at which a
u,v-component is induced or at which the loss of energy from the transmitter
occurs, allowing the resonance frequency to be measured.
It is possible to match the Larmor frequency and the applied radio frequency
either by holding the field strength constant (and hence a constant) and scanning
Sample
tu beA Am F.e Detector
Solenoid
H0 J---
H0
Hi
Radio
(A) transmitter (B)
FIGURE 7-11 Schematic diagram of a simple nmr spectrometer.

a variable applied radio frequency until matching occurs or, by varying the field
strength until o becomes equal to a fixed applied frequency. In the latter method,
fixed frequency probes (source and detector coils) are employed and the field
strength at which resonance occurs is measured. Two experimental configura-
tions are used in this experiment. In Fig. 7-1 l(A) the Ho field direction (z-
direction) is perpendicular to the sample tube, while in Fig. 7-11(B) the Ho field
and the sample are coaxial. There are some applications in which this difference
is important (vide infra).
We can now see why the relaxation processes discussed in equations (7-13)
and (7-14) had to be added to our complete equation ((7-20), which led to (7-25)
to (7-27) for the slow passage experiment) for the behavior of the magnetization.
If the populations of nuclei in the ground and excited states were equal, then
the probability that the nucleus would emit energy under the resonance condition
would equal the probability that the nucleus would absorb energy (i.e., transitions
mI(+1/2) -m 1(- 1/2) would be as probable as m(-'1/2) - m(+1/2). No net
change would then be detected by the radio-frequency probe. As mentioned
earlier, in a strong magnetic field there will be a slight excess of nuclei aligned
with the field (lower energy state) and consequently a net absorption of energy
results. In the absence of relaxation, energy is absorbed from the r.f. signal until
thepopulation in thilower state equals that in the higher state. This equilibrium
position is attained asymptotically. Initially, absorption might be detected but
this_absor-ption-wouldgradually disappear as the populations ofground and
excited states became equalFWhen this occurs, the sample is said to be saturated.
Without saturation, the relaxation mechanisms allow nuclei to return to the
lower energy state without emitting radiation. As a result there is always an
excess of nuclei in the lower energy state, and a continuous absorption of energy
from the r.f. source by the sample can occur.
The nmr experiment has significance to the chemist because nuclei exist in
atoms linked by bonds into molecules. The energy of the nuclear resonance (i.e.,
the field strength required to attain a Larmor frequency equal to the fixed
frequency) is dependent upon the electronic environment about the nucleus. The
electrons shield the nucleus, so that the magnitude of the field seen at the nucleus,
H,, is different from the applied field, H,:
HN = H,(1 - a) (7-28)
where a, the shielding constant, is a dimensionless quantity that represents the

shielding of the nucleus by the electrons. The value of the shielding constant
depends on several factors, which will be discussed in detail later.
In a slow sweep of H0 , the various magnetizations of the different nuclei
are sampled individually, because the nucleus is shielded or deshielded causing
HN to differ. For a molecule with two different kinds of hydrogen atoms this will
lead to a spectrum like that shown in Fig. 7-12.
One other point is worth making here. The differences in the agnetogric
ratios of different kinds of nuclei are much large than the-effectfrom , so there
isno trouble distinguishing signals from the different kinds of nuclei in a sample;
e.g., "F and 'H are never confused. The frequency ranges are so vastly separated
that different instrumentation is required to study different kinds of nuclei. In
Table 7-1, the resonance conditions (hv = gNflNH,) for various nuclei are given
for an applied field of 10,000 gauss.
Q.
a
.0
21
a)
C
w or H
FIGURE 7-12 A low resolution nmr spectrum of a sample containing

two different kinds of protons.
The frequencies in Table 7-1 are in MHz (106 Hertz or 106 c/s), and
the variation in the proton resonances of typical organic compounds from
different shielding constants is only 600 Hz at 10 kilogauss. In some paramag-
netic complexes, shifts of the order of magnitude of 840,000 Hz have been
observed; but even these could not be confused with a fluorine resonance. The
relative sensitivities of some nuclei in the nmr experiment are also listed in Table
7-1, with a more complete compilation given in Table 8-2.
The most common nmr experiment involves the uietho. We shall
discuss this experiment in more detail later, but now briefly consider it in the
context of equation (7-21). Suppose that a single strong H, pulse is imposed
(e.g., H, e 100 gauss > Ho - w,/ for all of the protons). Referring to the
equation for Hff.
Wi>
He, = H , + H,
TABLE 7-1. Important Nuclei in NMR
NMR
Frequency Relative"
in 10 Sensitivity Magnetic'
Abundance Kilogauss (constant Moment Spin'
Isotope (%) Fieldd HO) (p) (I)
1H
2
99.9844% 42.577 1.0000 2.7927 1/2
H (D) 0.0156 6.536 0.0096 0.8574 1
10B 18.83 4.575 0.0199 1.8006 3
"B 81.17 13.660 0.165 2.6880 3/2
13c 1.108 10.705 0.0159 0.7022 1/2
4 99.635 3.076 0.0010 0.4036 1
N
"N 0.365 4.315 0.0010 -0.2830 1/2
19F 100. 40.055 0.834 2.6273 1/2
29Si 4.70 8.460 0.0785 -0.5548 1/2
31P 100. 17.235 0.0664 1.1305 1/2
"?Sn 7.67 15.77 0.0453 -0.9949 '/2
119Sn 8.68 15.87 0.0518 -1.0409 '/2
For equal numbers of nuclei, where IH equals one.

In multiples of the nuclear magneton, eh/4nrMc.
cIn multiples of h/27r.
'MHz.
the local H, will vary slightly from nucleus to nucleus, but H1 is so large that
the term His, is completely dominant, and Her, is almost the same for all nuclei
present. All nuclei are sampled simultaneously. Since H1 is directed along u and
the net magnetization along z, the cross product requires that M begin to precess
about u. The pulse time is short, compared to the time for one full precession,
so the magnetization does not have time to precess around u, but merely tips
toward v. Accordingly, the v-component is measured in this experiment as 1M1 is
tipped toward v. Most experiments are now being done by Fourier transform
procedures. If a strong pulse were employed for a very long time, the nuclei
would precess around Herr as in the slow passage experiment; a u-component
would arise and v would disappear. Pulse experiments are not usually done this
way. When a strong pulse is employed for a short enough duration* that there
is no time for any relaxation to occur during the pulse (t, : 10 microseconds),
the Bloch equations predict the following expression for the u-component, as a
function of time, t:
M, = sin(7'Hit,)MO exp( -t/T 2) (7-29)
This expression results from the Fourier transform of equation (7-23), which
allows conversion from the frequency domain (7-23) to the time domain (7-29).
We shall discuss this topic in more detail later. The quantity dM,/dt cannot be
set equal to zero in this experiment. The v-component of the magnetization
decays with time according to equation (7-29), as shown in Fig. 7-13. This is
called a free induction decay curve and it can be analyzed by computer (via
Fourier transformation, vide infra) to give a frequency spectrum indentical to the
Lorentzian slow passage result. A comparison of the slow passage and Fourier
transform experiments is provided in Table 7-2. We shall cover fundamental
properties of nmr in this chapter that are independent of the way in which the
nmr experiment is carried out, i.e., slow passage or Fourier transform.
M,
'
Time
FIGURE 7-13 Free induction decay curve; a plot of the change

in magnetization of the sample with time.
* In Chapter 8, we shall provide an equivalent description of this experiment in terms of the

distribution of frequencies that make up the square wave produced from this pulsing.
TABLE 7-2. A Comparison of the Fourier Transform and the Slow-Passage

Experiment (100 MHz) in Terms of the Equation
yHera = (7 H 0 - 0 1 )2 ± yHNe1
Order of Magnitude
of Frequencies in
Experiment sec-1 (Hz) What Happens? What Is Measured?
yHo and o, ,Ho - UHb yH,
slow 108 100 100 ji's precess about

passage (at resonance) Hff many times as
we sweep through M
resonance, so M tips
toward u.
Fourier 108 104 106 pi's (and henceM )
transform precess about H,,
(FT) 1 of a revolution M
or less during the
pulse time, so M
tips toward v.
Multiplying by ,, puts Hff in units of frequency (Hz).
bIf you are accustomed to thinking in terms of parts per million, note that 104/108 100 x 10~'
100 ppm.
The Quantum 7-8 PROPERTIES OF I

Mechanical Now, to gain valuable insight, let us reexamine this whole problem using a
Description of quantum mechanical instead of a classical mechanical approach. Quantum
mechanics shows us that for I = /2, there are two allowed orientations of the spin
the NMR angular momentum vector in a magnetic field (Fig. 7-8) and will also indicate
Experiment the necessary requirements to induce transitions between these energy states by
an appropriate perturbation, i.e., the application of an oscillating magnetic field
with energies corresponding to r.f. radiation. The necessary direction for this field
can be determined from a consideration of the spin angular momentum oper-
ators.
Thql2 operator has eigenvalues I(I + 1) (as in an atom, where the orbital
angular momentum operator L 2 = I(/ + 1)h2 0). Any one of the components of
I (e.g., I.) commutes with P so we can specify eigenvalues of both P and [f.
The eigenvalues of fI are given by I, (I - 1), ...- I (like m, values in an atom,
where the z-component of angular momentum, L2, is given by Lzo = mhf with
m, = ±, ... , 0). In general, if two operatorscommute, then there exist simultaneous
eigenfunctions of both operatorsfor which eigenvalues can be specified.
It is a simple matter to determine whether two operators commute. If they
do, then by our earlier definition of commutation (Chapter 2)
I2f _ f P=0
The above difference is abbreviated by the symbol [P, I]. Similar equations can
be written for If and f,, i.e.,
[P, fj] = [p, Y] = 0

However, I does not commute with I, or I,, e.g.,
Ii, - i, =A
0
Eigenvalues for P exist, and if we decide to specify eigenvalues for I,, then
eigenvalues for I and f, do not exist.
Average values (Chapter 3) could be calculated for the I, and I, operators.
We shall make these concepts more specific by applying these operators to the
spin wave functions a and # for I = 2 nuclei and showing the results:
I, = (+'/2)0
'/2)#(7-30)
If x = (1/2)# fx# = (- |/2)1X

= (7-31)
Iya = (C/2)0 I,# = (- |/2)ia
Thus, the f, operator yields eigenvalues, since operation on a gives back a and
operation on # gives #l. The I and fI operators do not yield eigenvalues, since
operation on at produces # and operation on # yields 2. The average value for
2
the property fI or fI is given by an equation of the sort O*Opo dr/f 0 dr. The j
following relations hold for a and # (as they do for orthonormal electronic wave
functions):
f12 dT #2 d= 1 and fo/# dr = 0
As mentioned in Chapter 3, the integrals encountered in quantum mechan-

ical descriptions of systems are written employing the bra and ket notation for
convenience. Recall that the symbol ( I is referred to as a bra and | > as a ket.
An integral of the form j(O*
Op O)dz is written as (|rOpI0>, whereas an
j
integral of the form 1 1 *f 2 dr is written as (01| 1 2 >
7-9 TRANSITION PROBABILITIES
Consider the effect of the H 1 field in the quantum mechanical description. If the
alternating field is written in terms of an amplitude H, , we get a perturbing
term in the Hamiltonian of the form of equation (7-32):
Hpert = -yhH, I, cos w 1 t (7-32)
Recall from equation (7-16) that R for a nucleus in a z-field was

f = -yhHoI, = -gN3NHoI., so now the perturbation is of a similar form but
for an x-field that is alternating.
The equation describing the probability of a transition in the nmr, P, is
similar to that in u.v. and IR, and is given by
P = 2nyN2H12 2g(w) (7-33)

where g(w>) is a general line shape function, which is an empirical function that
describes how the absorption varies near resonance. To apply equation (7-33),
we need to evaluate matrix elements of the form <(pex|Ilp> and determine
whether they are zero or non-zero.
The solution is best accomplished by constructing a matrix that summarizes
all of the integrals that must be evaluated to describe the system in a magnetic
field with and without H1 . Rows and columns are constructed, which are headed
by the basis set. In this case, we have one nucleus with a and #l nuclear spin
wave functions leading to:
p
01 KaIH + Hpertk?> K<alH + HpV.tIB>
# (#I +f lpertL> (#|' + fl,.rl>
By this procedure, we have considered all possible matrix elements and also have
them in such a form that if E were subtracted from the diagonal elements we
would have the secular determinant, which can be solved to give the energies of
these states in an applied field (H, + H,).* The resulting energies could then be
used in the secular equations (produced by matrix multiplication of the secular
determinant with a matrix of the basis set) to give the wave functions in the field.
(Notice the analogy of this to our handling of the secular determinant and
equations in the section on Huckel calculations in Chapter 3.)
We begin by evaluating the elements in the secular determinant when the
applied field is H, (with a z-component only), i.e., the Zeeman experiment. Since
there is no x- or y-field component (only z), there are no Ix or I, operators and
all matrix elements of the form <fxj> or (IIjl are zero. The off-diagonal
elements, (alfI|#> and (#|fz|a> are also zero because fI#> = 1/ #, and 2
(a|#> and <#ix> are zero. The only non-zero matrix elements are <alfila> and
<B#|f|#>, with the former corresponding to stabilization, + '/2 (i.e., '/2(a la>), and
the latter to destabilization, - '/2. With no off-diagonal elements, the eigenvalues
are obtained directly and the basis set is not mixed, so the two wave functions
are a and #. When these are substituted into equation (7-33) for cp and pex, the
matrix element is zero with I, and the transition is not allowed, <|#> = 0.
Next, we shall consider what happens when an H, field along the x-axis
is added to the Zeeman experiment described above. We must now worry
about matrix elements involving f.. The diagonal elements <a|fIja> are zero
[(0 f4 10 = '/ 2 <a#I) = 0], but the off-diagonal elements (a[IfI#> are non-zero.
Since the H 1 field is small compared to Ho (z-component), these off-diagonal
matrix elements are so small as to have a negligible effect on the energies (the
effect of Iz on the diagonal is the same as in the Zeeman experiment). However,
the small off-diagonal matrix elements are important because they provide a
mechanism for inducing transitions from a to # because the new wave functions
for the system with H, present (i.e., obtained after diagonalizing our matrix) mix
a little # characterinto the a-Zeeman state and a little a into the #-Zeeman state.
The new wave function for the a-Zeeman state now is cp = /1 - a2|a> + alp),
* Recall that (OjfIt|> = E(010>. For an orthonormal basis set, (/'lfr,,,> equals I when n = m
but equals zero when n # m. Thus, E appears only on the diagonal of the energy determinant.
where a < 1, and a similar change occurs in the #l state. When these new wave
functions (p are substituted into equation (7-33), P is non-zero, making the
transition allowed. This corresponds to the classical picture of H, exerting a
torque to give a transverse component to the magnetization. Thus, we see that
the probability, P, of a transition occurring depends upon the off-diagonal matrix
elements in the 2,fl basis being non-zero, so that equation (7-33) is non-zero.
Next, consider what would happen if the r.f. perturbing field H1 was placed
along the z-axis. The off-diagonal matrix elements would again be zero, so
equation (7-33) becomes
P =(||#=0
This would correspond to a slight reinforcement of H0 , but would not allow

transitions. This exercise shows that H 1 cannot be colinear with H, to get an
nmr transition. (In the classical model, there has to be a perpendicular component
for H, to exert a torque.)
Finally, consider the case in which I = 1. Matrix elements <m' I,IIm> vanish
unless m' = m + 1. Consequently, for I = 1, the allowed transitions are between
adjacent levels with Am = ± 1, giving ho = AE = yhHo.
In summary, simply placing a sample in a magnetic field, H,, removes the
degeneracy of the m, states. Now, a radio-frequency source is needed to provide
hv to induce the transition. Absorption of energy occurs provided that the
magnetic vector of the oscillating electromagnetic field, Hi, has a component
perpendicular to the steady field, H0 , of the magnet. Otherwise (i.e., if H1 is
parallel to HO), the oscillating field simply modulates the applied field, slightly
changing the energy levels of the spin system, but no energy absorption occurs.
The influence of relaxation effects on nmr line shapes leads to some very Relaxation
important applications of nmr spectroscopy. Accordingly, it is worthwhile to Effects and
summarize and extend our understanding of these phenomena. We begin with
a discussion of relaxation processes and their effect on the shapes of the Mechanisms
resonance line. The lifetinte of a given spin state influences the spectral line width
via the Uncertainty Principle, which is given in equation (7-34):
AE At ; h (7-34)
Since AE = h Av and At = T2, the lifetime of the excited state, the range of
frequencies is given by Av 1/T2. The quantity 1/T2 as employed here lumps
together all of the factors influencing the line width (i.e., all the relaxation
processes) and is simply one-half the width of the spectral line at half-height.
When the only contribution to T2 is from spin-lattice effects, then T2 = T. Most
molecules contain magnetic nuclei; in the spin-lattice mechanism a local fluctuat-
ing field arising from the motion of magnetic nuclei in the lattice couples the
energy of the nuclear spin to other degrees of freedom in the sample, e.g.,
translational or rotational energy. The lattice refers to other atoms in the
molecule or other molecules, including the solvent. For liquids, T, values are
usually between 10-2 and 102 seconds but approach values of 10-' seconds if
228 Chapter 7 Nuclear Magnetic Resonance Spectroscopy--Elementary Aspects
certain paramagnetic ions are present. The extent of spin-lattice relaxation

depends upon (1) the magnitude of the local field and (2) the rate of fluctuation.
Paramagnetic ions have much more intense magnetic fields associated with-them
and are very efficient at causing relaxation. The water proton signal in a 0.1-M
solution of Mn(H 2 0) 6 2 is so extensively broadened by efficient relaxation from
Mn2 + that a proton signal cannot be detected in the nmr spectrum. As mentioned
before, the spin-lattice process can be described by a first order rate constant
(l/T,) for the decay of the z-component of the magnetization, say, after the field
is turned off, and is referred to as longitudinal relaxation.
Next, we shall discuss some processes that affect the xy-components of the
magnetization and are referred to as transverse relaxation processes. The spin-
lattice effect discussed above always contributes to randomization of the xy-
component; therefore,
1 1 1
T2 T, T2 '
where l/T 2' includes all effects other than the spin-lattice mechanism. When the
dominant relaxation mechanism is spin-lattice for both longitudinal and trans-
verse processes, i.e., 1/T1 = l/T 2, then l/T 2' is ignored. The quantity we used in
the Bloch equations is 1/T2 and not 1/T2'. In solution, these other effects, 1/T,',
are small for a proton compared to 1/T, so (1/T 2) = (1/T,). The other effects
include field inhomogeneity, spin-spin exchange, and the interaction between
nuclear moments. We shall discuss these in detail, beginning with field in-
homogeneity.
When the field is not homogeneous, protons of the same type in different
parts of the sample experience different fields and give rise to a distribution of
frequencies, as shown in Fig. 7-14. This causes a broad band. The effect of
inhomogeneity can be minimized by spinning the sample.
Interaction between nuclear moments is also included in T2'. When a
neighboring magnetic nucleus stays in a given relative position for a long time,
as in solids or viscous liquids, the local field felt by the proton has a zerofrequency
contribution; i.e., it is not a fluctuating field as in the T, process from the field
FIGURE 7-14 The same of the neighboring magnetic dipoles. A given type of proton could have neighbors
nmr transition occurring in a with, for instance, a + + - + - combination of nuclear moments in one
bulk sample in an
molecule, + - + - - in another, and so forth. Variability in the static field
inhomogeneous field.
experienced by different protons of the same type causes broadening just like
field inhomogeneity did. To give you some appreciation for the magnitude of
this effect, a proton, for example, creates a field of about 10 gauss when it is 1 A
away. This could cause a broadening of 10' Hz. As a result of this effect, extremely
broad lines are observed when the nmr spectra of solids are taken. This effect is
averaged to zero in non-viscous solutions by molecular rotation.
The process of spin-spin exchange contributes to T2 but not to T, for it
does not influence the z-component of the magnetization. In this process, a
nucleus in an excited state transfers its energy to a nucleus in the ground state.
The excited nucleus returns to the ground state in the process and simultaneously
converts the ground state nucleus to the excited state. No net change in the
z-component results, but the u- and v-components are randomized.
In common practice, l/T 2' is never really used. Either 1/T 2 = 1/T (with 1/T2 '
negligible) or we just discuss 1/T2 . In a typical nmr spectrum, 1/T2 is obtained
from the line shape (as will be shown) and it either equals 1/T, or it does not.
As mentioned earlier, the true form of a broadened line is described
empirically by a shape function g(w), which describes how the absorption of
energy varies near resonance according to equation (7-33). Since magnetic
resonance lines in solution have a Lorentzian line shape
2
g0)) - T2w (7-35)
7r I + T22(o) _ wo)
The width of the band between the points where absorption is half its maximum
height is 2/T 2 in units of radians sec-'. In units of hertz, the full bandwidth at
half height is given by 1/xT 2 .
7-10 MEASURING THE CHEMICAL SHIFT
The shielding constant, equation (7-28), makes nmr of interest to a chemist. In

the typical nmr spectrum, the magnetic field is plotted versus absorption as
illustrated by the low resolution spectrum of ethanol shown in Fig. 7-15. The
-CH 3
"CH2
OH
-- Increasing field strength (constant frequency)-

- increasing frequency (constant field)
- Increasing shielding
FIGURE 7-15 Low resolution proton nmr spectrum of C2 H5OH.
least shielded proton (smallest a) on the electronegative oxygen atom interacts

with the field at lowest applied field strength. The areas under the peaks are in
direct proportion to the numbers of equivalent hydrogens, 1:2: 3, on the
hydroxyl, methylene, and methyl groups. Note that the separation of absorption
peaks from -CH 3 and -CH 2- hydrogens in this spectrum is much greater
than that in the infrared spectrum.
We wish to calibrate the horizontal axis of Fig. 7-15 so that the field strength
(or some function of it) at which the protons absorb energy from the radio
frequency probe can be recorded. Equation (7-28) could be employed, but
accurate measurement of HN and Ho is difficult. Instead, a reference material is

employed, and the difference between the field strength at which the sample
nucleus absorbs and that at which the nucleus in the reference compound absorbs
is measured.
The reference compound is added to the sample (vide infra), so it must be
unreactive. Furthermore, it is convenient if its resonance is in a region that does
not overlap other resonances in the molecules typically studied. Tetramethylsi-
lane, TMS, has both properties and is a very common reference material for
non-aqueous solvents. In view of the limited solubility of TMS in water, salts of
the anion (CH 3)3 SiCD 2 CD 2 CO 2 ~ are commonly used. The position of its
resonance is set at zero on the chart paper.
The magnetic field differences indicated by the peaks in Fig. 7-15 are very
small, and it is difficult to construct a magnet that does not drift on this scale.
Accordingly, most instruments pick a resonance and electronically adjust the
field circuit to maintain or lock this peak at a constant position. This can be
accomplished by having some material in a sealed capillary in the sample tube
or in the instrument to lock on, or else by picking a resonance in the spectrum
of the solution being studied for the purpose. The former procedure is described
as an external lock and the latter as an internal lock. The internal lock produces
more accurate results. Temperature and bulk magnetic susceptibility effects are
minimized. The field is being locked on a resonance that has all of the advantages
of an internal standard (vide infra). In a typical experiment employing an internal
lock, the 2H signal of a deuterated solvent is utilized via a separate transmitter-
receiver system.
With TMS at zero, it is possible to measure the differences in peak maxima,
A. Although the field is being varied in the slow passage experiment, the abscissa
and hence the differences in peak positions are calibrated in a frequency unit of
cycles per second, referred to as hertz (Hz). This should cause no confusion,
for frequency and field strength are related by equation (7-5):
o= yH
According to equation (7-28), the shielding of the various nuclei, oHo,

depends upon the field strength, Ho. If a fixed frequency probe of 60 MHz is
employed, the field utilized will be different than if a 100-MHz probe is employed.
As shown in Fig. 7-16, the peak separations are 10/6 as large at 100 MHz as at
60 MHz
222 155 73 0
100 MHz
370 258 122 0
A, Hz (H increasing ---
FIGURE 7-16 The nmr spectrum of CH3CH20H at 60 Mhz and at 100 MHz.
60 MHz. To overcome this problem and to obtain values for the peak positions,
which are independent of field strength, the chemical shift, 6, is defined as
Ax 106
(7 36)
fixed frequency of the probe, v,
Since the probe frequency of v, is in units of megahertz (MHz), and A has units
of hertz, the fraction is multiplied by the factor 106 to give convenient numbers
for 6 in units of parts per million, ppm. The chemical shift, 6, is independent of
the probe frequency employed.
If the sample nuclei are deshielded relative to the reference peak, A is
positive* by convention. When Si(CH 3)4 is employed as a standard, most proton
6 values are positive and the larger positive numbers refer to lesser shielding.
For several reasons, it is advantageous to add the TMS directly to the
solution being studied as opposed to having it in a separate sealed capillary tube,
i.e., an internal vs. an external standard. Magnetic field-induced circulations of
the paired electron density in a molecule give rise to a magnetic moment that is
opposed to the applied field. In diamagnetic substances, this effect accounts for
the repulsion, or diamagnetism, experienced by these materials when placed in
a magnetic field. The magnitude of this effect varies in different substances, giving
rise to varying diamagnetic susceptibilities. This diamagnetic susceptibility in
turn gives rise to magnetic shielding of a molecule in a solvent, and is variable
for different solvents. The effect is referred to as the volume diamagnetic
susceptibility of the solvent. Thus, the chemical shift of a solute molecule in a
solvent will be influenced not only by shielding of electrons but also by the
volume diamagnetic susceptibility of the solvent. The 6 values obtained for a
solution and a pure liquid solute (this is often referred to as the neat liquid) relative
to an external standard (Fig. 7-17) would be different to the extent that the
volume diamagnetic susceptibilities of the solvent and neat liquid were different.
In solution, the diamagnetic contributions to the shielding of the solute depend
upon the average number of solvent and solute neighbors. Consequently, the
chemical shift will be concentration dependent. To get a meaningful value for 6
with an external standard, it is necessary to measure 6 at different concentrations
and extrapolate to infinite dilution to produce a value for 6 under volume
susceptibility conditions of the pure solvent.
Contributions to the measured 6 from the volume diamagnetic susceptibility
Sample
of the solvent can be more easily minimized by using an internal standard.The
internal standard must, of course, be unreactive with the solvent and sample.
Furthermore, the sample and standard must be uniformly dispersed. Under these - External
conditions the standard is subjected to the same volume susceptibility (from standard
solvent and solute molecules) as is the solute, and the effects will tend to cancel
when the difference, A, is calculated. (Due to variation in the arrangement of
FIGURE 7-17 Concentric
solvent around different solutes, an exact cancellation may not result.) For
nmr sample tube containing
accurate work, spectra at two different concentrations should be run as checks an external standard. For
and the results relative to two internal standards compared to insure that the illustrative purposes the
external standard
compartment size is
* In the early literature and even as late as 1970 in the area of "F and 31 P nmr, the convention exaggerated. This is really
used is the opposite of that described here. The reader must exercise considerable care in ascertaining just large enough for two to
the convention used. three drops of standard.
internal standard approximations are working. Molecular interactions of the

solvent with the standard or solute cause complications.
7-11 INTERPRETATION OF THE CHEMICAL SHIFT
a. Local Effects
The range of proton chemical shifts measured on pure liquids(4) for a series of
organic compounds is illustrated in Fig. 7-18. Proton shifts outside this total
range have been reported, and sometimes shifts outside the range indicated for
a given functional group occur. In general, the data in Fig. 7-18 serve to give a
fairly reliable means of distinguishing protons of functional groups. Note the
difference in CH 3 -C1, CH 3C=, CH 3-O, HC=, HCO, etc. Very extensive
compilations of proton chemical shifts have been reported,(5 ) which can be
employed for the fingerprint type of application. The reference section at the end
of this chapter lists compilations of shifts for various nuclei. The shift of OH
protons is very concentration and temperature dependent. For example, the
spectrum of ethanol changes as a function of concentration in an inert solvent
(e.g., CCl 4 ), by 5 ppm in going from concentrated to dilute solution. At infinite
dilution, the hydroxyl proton appears at a higher field than the methyl protons,
in contrast to the spectrum of pure ethanol represented in Fig. 7-16. There is
more hydrogen bonding in concentrated than in dilute solutions and hydrogen
bonding of the proton causes a shiftio a lower field. This behavior on dilution
can be iemployed to verify the assignment of a peak to a hydroxyl group, to
investigate the existence of steric effects in hydrogen bonding(6) and to distinguish
between intermolecular and intramolecular hydrogen bonding. Solvent effects
are quite large whenever specific interactions occur and, as will be shown later,
nmr has been very valuable in establishing the existence or absence of these
interactions. The differential shift of the C-H and 0-H protons of methanol
or ethylene glycol is the basis of a standard technique for estimating probe
temperature. 7 )
The interpretation of proton chemical shifts is complicated by the existence
of contributions from local (i.e., the atom undergoing the transition) shielding
and neighbor atom anisotropic effects. Local shielding arises from magnetic field-
induced electron circulation on the atom undergoing the transition. This shield-
ing is a tensor quantity with u, giving the contribution parallel to the field when
the z-axis of the molecule is aligned with the field. Using perturbation theory
and treating the magnetic field as a perturbation on the ground-state molecular
wave function, 10>, leads to the Ramsey equation(8):
" 2mnc 2 K r3
(diamagnetic term)
(e
2~ OIzlKl2L
<0|L,|n) n -- 2 0) ) (0K 2L7nKnL~
-| n(||0
r + (7-37)
2mc,/ , E-E E, -E,
(paramagnetic term)
7-11 Interpretation of the Chemical Shift
f (PPM TO TMS)
16 14 12 t0 a 6 4 2 0
TMS
-C02 2. (20
0
-Ct (29)
C-NOH 15)
N-CH (4)
A H100)
RC-NR(H) 28)
A,0H (76)
C.CH(.200)
C=C-Nd(27)
O C'C (161
ROH (29)
RNH (11)
C - 24) FIGURE 7-18 Hydrogen

R. - CH3-H CH ,-C'-C
C
CH2 - (1400)
nuclear magnetic resonance
CH3 - (13001
TMS chemical shift correlation
I . I I I . I I I . I I I . charts, (A) for main proton
16 14 12 10 8 6 4 2 0
groups, (B) for CH, OH, and
(A) J (PPM TO TMS)
NH subgroups. Open bar
denotes extreme 10% of
d (PPM TO TMS) data. [from M. W. Dietrich
7 6 5 4 3 2 1 0 and R. E. Keller, Anal.
TMS Chem., 36, 258 (1964).]
Illustration continued on
E3M ArOCt 05)
following page.
POCH3 (12)
SOCH3 (6)
0
OCM
E~: 3
222
01001 R0CM3 (1
1
C-NCt. (6)
A, NCH3 (42
N RNCtj3(7)
OsCH 3 AND CSCN (101
3
I l Y- C-CH 3 (20)
. M Ar(R)(H)- CH3 (26)
E Z ACH 3 (39)
C C=C-CH3AND C!C-CL43 (681
R- -CH 3
. -CH.~~ ,_-C ~ t 3 A0~C 3
C C
G PCM3 (9)
Y3C-CH3AND (0)
Y2RC-CH
3
ORGANIC
y.*COMMON STRUCTURAL YRCHC3 (43)
OTHER THAN R
GROUPS YCH2 CI(3 (41)
RC43 (37)
TIS
7 6 D 4 2 1
(B) jd (PPM TO TMSP
((PPM TO TMS)
I3 6 94 12 10 S 6 4 2 0
I ' ' I ' I ' I ' I ' I
TMS
0
A, Nt4c (4)
(C-C2NH
(2 0
RCNHR(H)(18)
A,NH (15)
RNH (10)
FIGURE 7-18 (continued) -0(

C=C'C 0 (6)
Ar0H4(76)
cc
6o (30)
R. -CH3,-C 2 -d - CYYY"8)
CHRYY' (36)
Y. COMMONORGANICSTRUCTURAL (59)
CHRR'Y
GROUPSOTHER THAN R [ CHRR' f91
TMS
,. , | I 1, 1.. 1 Il . ITUS|
12 8 6 4 2 0
Is 16 14 10
(C) -((PPM TO TMS)
This equation is similar to the Lamb equation for an atom, and it tells us what
the applied field does to our molecule. A positiyQ. sigmia.ts shieldingie., an
upfieldora dimagnetic contribution. The negative sign indicates deshielding,
i.e., a downfiel ora param" on.
The first term on the right-hand side of the equality is the so-called
diamagnetic term. The symbol <01 corresponds to the ground state wave function,
and r is the distance from the electron to the nucleus undergoing the transition.
Since only <01, the ground state wave function, is involved in this first term, no
excited states are mixed in by this term. The field does not distort the electron
distribution in the molecule, but just induces a spherical electron circulation. If
this were the only effect, the molecular wave function would be independent of
the magnetic field.
The matrix element corresponding to this first term reduces to:
(e2)2mc2)f (0 1 2 /r)dT (7-38)
The term 2 refers to the effective number of electrons in the s-orbital. This effect
gives rise to the normal diamagnetism observed for S = 0 molecules and can be
pictured as the shielding of the nucleus resulting from field-induced circulation
of electron density as illustrated in Fig. 7-19.
The field generated at the nucleus by electron circulations is directly
proportional to the strength of the applied field and will also depend on the
electron density surrounding the nucleus. If this were the only effect of the
magnetic field on a molecule, the chemical shift of a hydrogen nucleus would
parallel the electronegativity of the groups attached. This correlation is not
observed because of the neighboring anisotropic contributions to be discussed
shortly.
For simplicity, we have treated only the o,, component to illustrate the
factors that contribute to the shielding. There are similar expressions for a., and
Electron
circulation
FIGURE 7-19 Local diamagnetic

contribution to the shielding. Electron
LN circulations occur in a plane
perpendicular to the plane of the page.
The induced field given by the arrow in
the middle is opposed to HO.
Lines of magnetic flux
a,, which, in general, are not necessarily equal to a., (for a., the matrix element
is <01(y 2 + z 2 )/r3 10>) and the three quantities must be averaged to produce the
shielding observed when the molecule is rapidly tumbling
S=(ax + a,, + Oa,) (7-39)

3
For a hydrogen atom, the three contributions are equal for this local effect.
The range of chemical shifts for nuclei other than hydrogen is considerably
larger. For example, the difference between the fluorine resonances in F2 and in
Fa is 542 ppm, compared to the range of about 12 ppm for proton shifts (see
Fig. 7-18). A brief compilation of fluorine shifts is contained in Table 7-3 to
illustrate this point. A wide range (- 500 ppm) of phosphorus chemical shifts and
"C shifts are also observed. References are given at the end of this chapter for
compilations. The fingerprint application is immediately obvious for compounds
containing these elements.
The large range in chemical shifts for these nuclei arises from a local effect
corresponding to the second term of the Ramsey equation. This is called the
paramagnetic term.* We are not using the word paramagnetic here to connote
the same thing as when it is used in connection with molecules containing
TABLE 7-3. Fluorine Chemical Shifts of Selected Compounds (in ppm Rela-
tive to CFCI 3 -Larger Negative Numbers Indicate Higher Field)
CH 3F -272 CF 4 -70 SOF 2 +70

HF -203 TeF, -64 TiF 6 2- +75
(CF 3 )3CF -188 (CF 3)4C -61.5 CIF3 +80
CF 6 -163 AsF 3 -48 BrF, +132, +269
SiF 4 -177 BrF 3 -38 NF 3 +140
BeF2 -155 CF 3CI -32 CF 30F +142
BF3OEt 2 -153 CF2 Cl2 -9 SFOF +178
BF4 - -149 CF3 I -4 FOClO 3 +225
BF, -131 CFC13 0 IF, +238, +274
SiF 6 2- -127 IF, +4 OF 2 +250
(CF 3CF 2)2 -127 S0 2F2 +36 FC1O 2 +288
F -(aq) -120 SF, +42 FNO 2 +394
SbF 6 - -109 SFOF +48 F2 +422
SbF, -108 SeF, +55 XeF 4 +438
CF 3H -88 IF, +53 FNO +478
SbF 3 -86 N 2 F4 +60 XeF 6 +550
(CF 3)2CO -82 NSF 3 +66 UF6 +746
CF3COOH -77
unpaired electrons. It is used here to describe the contributions from the

field-induced nonspherical circulation of the electron density, i.e., the second term
on the right-hand side of the Ramsey equation. The term involves field-induced
mixing in of excited states with the ground state, and this gives rise to a
mechanism for non-spherical electron circulation and the accompanying para-
magnetic contribution to the shielding. The equation for the paramagnetic effect
corresponds very closely to the paramagnetic term in the Van Vleck equation
for magnetic susceptibility (vide infra). The following quantities appear: <01
ground state and (ni excited state wave functions; L., the orbital angular
momentum operator; and E, - E0 , the energy difference between the ground
state and the excited state being mixed in. The summation is carried out over
all excited states; thus, to use the equation one needs wave functions and energies
of all excited states, including those in the continuum. The energies of virtual
orbitals (empty ones) are difficult to calculate, and thus it becomes impossible
to employ the Ramsey equation in the rigorous calculation of the chemical shift
of most molecules.
Evaluation of the entire second term of equation (7-37) and replacement of
the E. - EO term in the denominator by an average energy for the excited states
in the molecule (i.e., use of some electronic transition in the molecule that is felt
to be an average of all possible transitions to excited states) yields the para-
magnetic contribution, up, by averaging ax, og, and a,-:
2 (eh) 2
1 1 4I
up = (7-40)
3 mc r3 AE
The average energy approximation is often referred to as the closure approxima-

tion. Substituting appropriate values for F 2 into equation (7-40) (i.e., the energy
of 4.3 eV for the 7 -+ a* transition and <1/r 3> for a fluorine atom 2p orbital) one
obtains a, 2000 ppm. Since F has spherical symmetry, there can be no
angular momentum associated with the electron density in this species, and all
the paramagnetic terms in the Ramsey equation must be zero. A chemical shift
difference between F and F 2 of about 2000 ppm is expected compared to only
542 ppm observed. However, it is impossible to obtain a fluoride ion that is not
solvated or ion paired. The F2 chemical shift is found to be 625 ppm downfield
from HF. This is due in part to the poor nature of the average energy
approximation and in part to the use of an atomic <1/r 3> value. however, in
spite of the poor quantitative calculation of u,, there are several important
generalizations that can be drawn from this discussion. Since L is zero for an
s orbital in hydrogen compounds, one must invoke transitions to the 2p orbital
in order to obtain a paramagnetic contribution. Now the E, - E, value is so
large that paramagnetic contributions are not very important in the proton nmr
spectrum. These paramagnetic contributions become very large when both an
asymmetric distribution of p and d electrons in the molecule and low-lying excited
* The distinction between diamagnetic and paramagnetic contributions in equation (8-31) is

somewhat artificial, for it depends upon the gauge selected for the vector potential (see Slichter text.
p. 76, under Additional References). The traditional gauge corresponds to the measurement of angular
momentum about the nucleus whose chemical shift is being calculated. The total shielding is gauge
invariant.
states exist. The paramagnetic term gives rise to the principal contribution to
the shift in 9 F and "C nmr, and a large chemical shift range is observed. The
chemical shifts in "B nmr have significant contributions from both local
paramagnetic and local diamagnetic effects.
There is one additional bit of insight provided by the second term of the
Ramsey equation that will concern us. Since this term is important in atoms
having accessible p orbitals, our concern will be with the behavior of the L
operator on these orbitals. The behavior of L. is summarized as follows:
Llpz) = m pz> = Olpz> (7-41a)
Lzj p ) = i~p,) (7-41b)

Lzpy)= -iipx> (7-41c)
The results for the E, and L, operators will be presented as needed. We see from
equation (7-37) that we need to evaluate matrix elements of the type
(7-42)
0 -- n ) (nL 0>
r3
where : = z, x, and y for the azz, a , and cr,, components of the shielding tensor
for all possible excitations of an electron from the ground state m.o.'s to U2p-*.
(To simplify the problem, higher excited states will be omitted because the
E - E, term becomes very large for molecular orbitals derived from 3s, 3p, and
higher atomic orbitals.) We shall illustrate the evaluation of these matrix elements
for the F 2 molecule (molecular orbital description, valence orbitals
4
U2s 2s*262, (7r = ,) (rX,* = r,*)402p,*)
To be systematic, we shall consider the three different orientations in which the

molecular axes are aligned with the z-axis parallel to the field, with the x-axis
parallel to the field, and with the y-axis parallel to the field to evaluate a., (7 ,
and c,,, respectively. This requires the use of the Lz, L, and L, operators,
respectively, in the Ramsey equation.
Consider the cases where:
1. The z-axis of the molecule is aligned with the field. When the z-axis
of the molecule is parallel to the applied field, the molecule senses no H, or H,
field, so there are no field-induced L, or L, components. All one-electron
excitations place an electron into the antibonding molecular orbital, leading to
matrix elements such as:
(<2 Lj*),K ~p1LJU*>

(,7r1j a*> or (rtLz.r*>
where (6 2pLj Q0*> = ( / 2 )((Pza + pzb)LzRpza - pz,)> for a bond between fluorine
atoms A and B evaluating the contribution at A. We know that LIpz> = 0. Thus,
all the matrix elements for this orientation of the molecule are zero, and o is
zero. This leads to a generalization that will become important in subsequent
discussion of the chemical shift. Namely, the contribution to the chemical shift
from paramagnetic terms, a,, is zero when the highest-fold symmetry axis (z-axis)
is parallel to the fe7d.
2. The x-axis of the molecule is aligned with the field. That is, the z-axis
is perpendicular to the field. For the x-axis aligned this corresponds to evaluating
axx, and the L, operator. Considering one of the possible transitions, we have
(o-|L,|a*> = (1/ 2 )KPza|Lxjpza> + pza|ExIpr - (PzbILxPza>K

- Pzblxprb>]
Since the last term involves the L. operator centered on a operating on b, integrals
of this sort will generally be small; they are dropped here and in subsequent
discussion. The (pzPbxIpza> and (pzaILxIPzb> terms are two-center integrals;
they, too, generally are small and will not be considered in future discussion. In
the case under consideration, all of these matrix elements are zero because L.Ipz>
equals ilp,>, so:
(pz|L\pz> = <pjip > = 0
Only those matrix elements corresponding to (px|LIpz> can be non-zero

and these arise from
(r\Lxju*> and <*\L,\a*>
There is only one such matrix element to evaluate, namely, ('/2)(pxa|Lxlpza>

(recall that <pxaILpz)b> is a two-center integral and is small in comparison),
which is equal to i/2. This value corresponds to the matrix element (K|Lla*>.
The element <7r*ILla*> gives the same result. Thus, there is a contribution to the
paramagnetic term for the field along the x-axis. We see that the field has coupled
the ground state with the excited state; i.e., we have field-induced mixing.
3. The y-axis of the molecule is aligned with the field. This gives the
same result. We must now concern ourselves with L, in the molecular coordinate
system. The result is
L,|pz )= ilp,)
By analogy with our discussion of L, only the matrix element (pyaLylpza> is

non-zero, and we also expect a paramagnetic contribution to the shielding at
fluorine when the field is along the y-axis. As before, the other matrix elements
are zero.
b. Remote Effects
Protons attached to metal ions are in general very highly shielded,") the
resonance often occurring 5 to 15 ppm to the high field side of TMS and, in
some cases, occurring over 60 ppm upfield from TMS. A local contribution
cannot explain these large shifts. The existence of a field-induced moment on a
neighboring atom can be felt at the atom whose nmr is being studied. These are
referred to as remote effects or as neighbor anisotropiccontributions. The field at
this remote atom may be dominated by either a paramagnetic or a diamagnetic
effect but may make a different contribution at the atom being studied. We shall
work through the case of the HX molecule first, assuming that the diamagnetism
of the remote atom X is dominant. As illustrated in Fig. 7-20, the field at the
HH
\/x 1 /
(A) (B)
FIGURE 7-20 Neighbor anisotropy in HX for different orientations.
proton in HX from the diamagnetic effect at X will be strongly dependent upon

the orientation of the HX molecule with respect to the direction of the applied
field, H. When the applied field is parallel to the internuclear axis, the magnetic
field generated from diamagnetic electron circulations on X (indicated by the
dotted lines) will shield the proton (Fig. 7-20(A)), while in a perpendicular
orientation (Fig. 7-20(B)) this same effect at X will result in deshielding at the
proton. The magnitude of the induced moment on atom X (and hence the field
at the proton from this neighbor effect) for the parallel and perpendicular
orientations will depend upon the susceptibility of X for parallel and perpendicu-
lar orientations, xe and Z/ respectively. The susceptibility, X, is related to the
intensity of the magnetization, M, by
M
y = (7-43)
H0
When HX is parallel to the field, the induced moment on X is given by ZllxHO.

The contribution this remote effect makes to the shielding at the proton, al , is
given by -a1 HO; expressed in terms of the susceptibility of X, al is:
a1 = - 2R -3 Zx) (7-44)
where R is the distance from X to the proton, Z is negative (and consequently

the proton is shielded), and Zl, is the parallel component of the susceptibility
at X.
For the perpendicular (1) orientation of HX, the contribution from X at
the proton is given by
ai = R 3(X) (7-45)
Note that equations (7-44) and (7-45) give the correct signs for the shielding at
the proton, as illustrated in Fig. 7-20 (xZ and Z, are both negative). We have
two perpendicular components to consider, Z. and Z,, which are equal here and
in any molecule with a threefold or higher symmetry axis. In solution, the
molecule is rapidly tumbling, so the concern is with the average value of a given
by equation (7-46):
a= -3'R - 3(2 2Z,) (7-46)

According to equation (7-46), if the susceptibility is isotropic (i.e., X1 = xi) there

will be no contribution to the shielding at the nucleus of interest from diamagnetic
effects on the neighbor. In CH 4 , for example, Z1 = X, and neither the carbon
nor the C-H bonds can make a neighbor contribution to the proton shift. In
HX molecules, the remote contribution from X arises because I, # xZ.The value
to use for R in equations (7-45) and (7-46) is a problem, for one must decide
whether the susceptibility arises on the remote atom or in the bond. This question
is not a real one but results from our arbitrary factoring of the molecular
susceptibility into parts due to different atoms in the molecule. In spite of this
simplification, the approach is valuable and one often selects the atom or center
of the bond as the point from which to measure R.
If one had a molecule in which paramagnetic contributions were dominant,
the same equations would be employed, except that the signs of x would be
positive; i.e., the boldface arrow on X of Fig. 7-20 would point in the opposite
direction. Paramagnetic contributions from X (or the H-X bond) will be zero
for the orientation in which the bond axis is parallel to the field and will be a
maximum for the perpendicular orientation. Thus, the paramagnetic term will
be very anisotropic and often dominates any neighbor anisotropic contributions
to the proton shifts. Refer to Fig. 7-20(B) and change the direction of the arrows
for this paramagnetic effect at X. The paramagnetic effect at X is shielding
(diamagnetic) at the proton. When discussing a remote effect, we label it
according to the direction of the effect on the remote atom. If x1 and Z, were
known, we would know its sign and which effect is dominant. This information
is seldom available and both paramagnetic and diamagnetic neighbor effects
must be qualitatively considered when interpreting proton shifts.
Equations (7-44) and (7-45) apply only for linear molecules. For the general
case
a= ('1 3)R ~'[(1 - 3 cos 2 O8)xxx + (1 - 3 cos 2 O,)Zy, + (1 - 3 cos

2
0z)xzz] (7-47)
where x.., y, and Z., are the values along the three principal axes of the
susceptibility tensor and 0., 8,, and 02 correspond to the angles between these
axes and a line drawn from the center of the anisotropic contributor (the neighbor
atom or bond) to the atom being investigated. The angle is illustrated for a
molecule of C3, symmetry in Fig. 7-21, where the source of the neighbor effect
has been taken at the iodine atom, z., is taken along the threefold axis, and
xx = 7,,. When xxx # x,, # 7,2, we can select our coordinate system such that
becomes
R the radius vector R lies in the yz- (or xz)-plane. Equation (7-47) then
a = iR -3(2AxI - Ax2 - AZ 1 3 cos 2 O) (7-48)

C .
H
H '' where AZ1 = x2, - ,, and Ax 2 = zz - Xxx. When axial symmetry pertains,
Z22 # xx = x,, and equation (7-48) becomes (7-49):
H
FIGURE 7-21 Illustration of ~= (3)R - 3(Z - )(1 - 3 cos 2 = (11/)3)R - 3 A 7 (1 - 3 cos 2 g ) (7 49)
the parameters 0 and R of
equation (8-37). Values for AZ for various bonds can be found in the literature.
7-13 Examples of Chemical Shift Interpretation 241
7-12 INTERATOMIC RING CURRENTS
Interatomic ring currents develop in cyclic, conjugated systems. Field-induced

electron circulations occur in a loop around the ring and extend over a number
of atoms. Analogous to the circulation of electrons in a wire, a magnetic moment
is induced by the effect. The moment induced at the center of the ring is opposed
to the field, but in benzene, for example, the magnetic lines of flux at the protons
are parallel to the applied field, and these protons are deshielded as shown in
Fig. 7-22. For another orientation of benzene relative to the field, the ring current
shifts the protons to a smaller extent. The average overall orientation is
deshielding. This is the explanation for the large deshielding observed for the
protons in benzene."" Johnson and Bovey"') have calculated the ring current
contribution to the chemical shift for a proton located at any position relative
to the benzene ring. The acidic hydrogen, upon hydrogen bonding to benzene,
is shielded, whereas deshielding is observed with most donors other than benzene.
This indicates that the proton of the Lewis acid is located on the C, axis of
benzene.
7-13 EXAMPLES OF CHEMICAL SHIFT

INTERPRETATION
One important generalization can be drawn from the preceding and subsequent
discussion, viz., chemical shift data are not reliable indications of the electron
density around the nucleus being measured. Many effects contribute to 6.
Contributions from neighbor anisotropy are qualitatively invoked when needed
to account for differences between measured 6 values and those expected on the
basis of chemical intuition. The molecule is examined to see what property it
possesses that could account for the observed discrepancies. For example, the 6
value for HCl (in the vapor phase) indicates that this proton is more shielded
than those in methane. This is in contrast to the greater formal positive charge
on the proton of HCl than on that of CH 4 . Thus, the local diamagnetic effect
does not explain this behavior. The local paramagnetic effect is not expected to
be significant for a proton. Since HCl is linear and cylindrically symmetric about
the hydrogen-chlorine bond, a neighbor paramagnetic contribution is expected
when the molecule is perpendicular to the applied field but not when it is parallel
to the field. The net effect will be a shielding of the proton in HCl from a
Ho
FIGURE 7-22 Ring currents in benzene.

paramagnetic, neighbor anisotropic effect. The effect is paramagnetic (deshield-

ing) at X but shielding at the proton. The neighboring anisotropic effect
dominates the difference in CH 4 and HCl shifts.
In the series HF, HCl, HBr, and HI, the magnitude of the neighboring
paramagnetic contribution, Aou, giving rise to anisotropy increases with increas-
ing atomic number of the halogen. This parallels the decreased difference in the
energy of the ground and excited states, AE, making a field-induced mixing of
ground and excited states in the compounds of the higher atomic number atoms
more favorable. The Ramsey equation states that the paramagnetic contribution
to Au will be proportional to -1 /AE.
The formal positive charge on the proton for the compounds in the series
C2 H, < C 2 H4 < C 2 H 2 increases in the order listed. The 3 values increase in the
H C =C-- H order C2 H, < C2 H2 < C2 H4 , indicating decreased shielding. Acetylene is more
highly shielded than is expected on the basis of its acidity, and the shielding is
attributed to two effects: (1) Remote diamagnetic shielding from electron circula-
tions in the triple bond (Fig. 7-23(A)). (This effect is a maximum when the
(B) molecular axis is aligned with the field, and it is classified here as a diamagnetic
effect because the moment arising from the electron circulations opposes the
FIGURE 7-23 Contributions applied field. Recall that the paramagnetic contribution is zero for this orienta-
to the shielding in acetylene. tion and is a maximum when the axis is perpendicular to the field.) (2) Remote
(A) The effect here is a paramagnetic shielding from higher state mixing (Fig. 7-23(B)). Both the remote
diamagnetic effect at carbon, diamagnetic and paramagnetic effects influence AX so as to provide shielding at
which is a maximum for this the proton. The measured proton chemical shift of acetylene relative to ethane
orientation. (B) The effect is thus explained by the remote diamagnetic and paramagnetic contributions
here is a paramagnetic effect
arising from the triple bond.
at carbon, which is a
maximum for this orientation. In addition to the trends in the population of the hydrogen Is orbital as a
consequence of the electronegativity of the attached group, Buckingham has
proposed" 1 electricfield and reaction field models to account for changes in the
electron density of a bound hydrogen atom.
Suffice it to say that in view of the many contributions described above,
even the qualitative interpretation of proton chemical shifts is tenuous. When "C
and 9 F chemical shifts are considered, the dominance of the local paramagnetic
term gives rise to such large shift differences in series of molecules that neighbor
anisotropy effects become small in comparison." The qualitative interpretation
of the shifts in these compounds has met with more success." 1 The problem is
still not simple by any means. Changes in the average excitation energy,
one-center terms, and two-center terms in a molecular orbital description
are all important, 15' 16) leading to the situation where there are many more
parameters influencing the measurement than there are experimental observables.
The reader is referred to references 13, 15, and 16 in the event one is ever tempted
to propose a rationalization of chemical shift data.
The chemical shifts of analogous atoms in two different enantiomers are
identical. However, as with many physical properties, differences are detected in
diastereoisomers. Advantage has been taken of this fact"' 1 to utilize nmr for the
determination of optical purity. For example, when a dI mixture of the phosphine
0CH,
oxide, C 6HCH,-P is added to the enantiomeric solvent d-2,2,2-
CH 3 5 H I-
|pCFe3) th hy r g nb n e ad ut a e
trifluoro-1-phenyl ethanol (C6 H - O )the hydrogen bonded adducts are
diastereoisomers in which the chemical shifts of the methyl groups differ by
1.4 Hz. The difference in shift arises from differences in the neighbor anisotropic
contributions in the diastereoisomers. These peaks can be integrated and the
relative amounts of the d,l isomers in the phosphine oxide determined. The optical
purity of a resolved product can be ascertained by the absence of one of the peaks.
7-14 EFFECT OF SPIN-SPIN SPLITTING Spin-Spin

ON THE SPECTRUM Splitting
When an nmr spectrum is examined under high resolution, considerable fine
structure is often observed. The difference in the high and low resolution spectra
of ethanol can be seen by comparing Fig. 7-24 with Fig. 7-15.
The chemical shift of the CH 2 group relative to the CH 3 group in Fig. 7-24
is indicated by A measured from the band centers. The fine structure in the
CH 3 and CH 2 peaks arises from the phenomenon known as spin-spin splitting,
and the separation, J, between the peaks comprising the fine structure is referred
to as the spin-spin coupling constant. J is usually expressed in hertz, and typically
JHCCH is about 7 Hz and JHOCH about 5.5 Hz. As mentioned earlier, the magnitude
of A depends upon the applied field strength. However, the magnitude of the
spin-spin coupling constant in Hz is field independent.
FIGURE 7-24 High resolution nmr spectrum of ethanol

(facsimile). Compare with Fig. 7-15, p. 229.
l 4 -JH-C-C-H rH-C-CH
JHOCH JHOCH H I1 I f I
FIGURE 7-25 The interpretation in terms of the appropriate coupling constants of
the spectrum in Fig. 7-24. The meaning of this stick-type spectrum will be made
clear shortly in the text.
The cause of the fine structure and the reason for the field-independent
character of J can be understood by considering an H-D molecule. If by some
mechanism the magnetic moment of the proton can be transmitted to the
deuteron, the field strength at which the deuteron precesses at the probe
frequency will depend upon the magnetic quantum number of the neighboring
hydrogen nucleus. If the proton nucleus has a spin of + 1/2, its magnetic
moment is aligned with the field so that the field experienced by the deuterium
is the sum of the proton and applied fields. A lower applied field strength, H0 ,
will be required to attain the precession frequency of the deuterium nucleus in this
molecule than in the one in which the hydrogen has a magnetic quantum number
of -'/2 In the latter case the field from the proton opposes the applied field and
must be overcome by the applied field to attain a precessional frequency equal
to the probe frequency. The simulated spectrum that would result for the
deuterium resonance is indicated in Fig. 7-26(A).The m, values for the hydrogen
nuclei in the different molecules that give rise to different peaks are indicated
above the respective peaks. The two peaks are of equal intensity because there
is practically equal probability that the hydrogen will have + 1/2 or -'/2
magnetic quantum numbers. The proton resonance spectrum is indicated in Fig.
7-26(B). These proton resonance peaks correspond to magnetic quantum num-
bers of + 1, 0, and -1 for the attached deuterium nuclei in different molecules.
The spin-spin coupling constants JDH and JHD in Fig. 7-26(A) and 7-26(B),
respectively, have the same value. Subsequently, we shall discuss mechanisms for
transmitting the magnetic moment of a neighboring atom to the nucleus
undergoing resonance.
Returning to ethyl alcohol, we shall examine the splitting of the methyl
protons by the methylene protons. The two equivalent protons on the CH
2
group can have the various possible combinations of nuclear orientations
indicated by the arrows in Fig. 7-27(A). In case 1, both nuclei have m, values of
+2, giving a sum of +1 and accounting for the low field peak of the CH 3
resonance. Case 2 is the combination of CH 2 nuclear spins that gives rise
to the middle peak, and case 3 causes the high field peak. The probability that
the spins of both nuclei will cancel (case 2) is twice as great as that of either of
the combinations represented by case 1 and case 3. (There are equal numbers of
+ 1/2 and - '/2 spins.) As a result, the area of the central peak will be twice that of
the others (see Fig. 7-24).
The nuclear configurations of the CH 3 group that cause splitting of the
CH 2 group are indicated in Fig. 7-27(B). The four different total net spins
give rise to four peaks corresponding to the larger separation in this multiplet,
shown in Fig. 7-24. The relative areas are in the ratio 1: 3: 3:1. The separation
between these peaks in units of Hz is referred to as JH-C-C-H or as 3 H-H- The
superscript in the latter symbol indicates H-H coupling between three bonds.
The peak separation in the methylene resonance from this coupling is equal to
the peak separation in the methyl group. The spectrum of the CH 2 resonance is
further complicated by the fact that each of the peaks in the quartet from the
methyl splitting is further split into a smaller doublet by the hydroxyl proton.
In the actual spectrum, some of the eight peaks expected from this splitting
overlap, so they are not all clearly seen.
FIGURE 7-26 NMR spectra 2 +1 0 1

for the hypothetical HD
experiment. (A) Deuterium
resonance for HD; (B) proton
resonance for HD.
Inc. field strength - Inc. field strength

(A) (B)
Case IM,
1 +1
0
--1
FIGURE 7-27 Possible

Case orientations of proton
1' +11 nuclear magnetic moments
+12 for (A) -CH 2- and (B)
2'
-~
-~
-5
~-
*-
--- 5 +2 -CH 3 groups.
~- -~ -~ 2
3'
2
4'
2
The OH peak is split into a triplet by the CH 2 protons with the same
separation as HH--C-OH in the methylene resonance. Usually, though not always,
the effects of spin-spin coupling are not seen over more than three sigma bonds.
Accordingly, the interaction of the OH proton with the methyl group is not
seen. The spectral interpretation is illustrated in Fig. 7-25 by showing which
lines arise from which couplings. This "stick-type spectrum" is constructed by
drawing a line for each chemically shifted different nucleus. On the next row, the
effect of the largest J is shown. Additional lines are added for each J until the
final spectrum is obtained.
Because of the selection rules for this process (vide infra), equivalent nuclei
do not split each other; e.g., one of the protons in the CH 3 group cannot be
split by the other two protons. A general rule for determining the splitting pattern
can be formulated that eliminates the necessity for going through a procedure
such as that in Fig. 7-27. For the general case of the peak from nucleus A being
split by a non-equivalent nucleus B, the number of peaks, n, in the spectrum of
A is given by the formula
nA= 2
SB + 1 (7-50)
TABLE 7-4. Spin-Spin Splitting in Various Moleculesa
Groups Being Groups Doing

Molecule Split (A) the Splitting (B) ESB n
CH 3 CH 2 OH CH 3 CH 2 1 3
CH 3 CH 2 OH CH 2 CH 3 3/2 4
PF 3 P F 3/2 4
PF 3 F P 1/2 2
(CH 3 )4 N* CH 3 N 1 3
4 Spin quantum numbers for the most abundant isotopes of nuclei con-
tained in the above compounds are H = '/2, P = /2, F = '/2, N = .
where ESB equals the sum of the spins of equivalent B nuclei. The application
of this formula is illustrated by the examples in Table 7-4. The relative intensities
of the peaks can be obtained from the coefficients of the terms that result from
the binomial expansion of (r + 1)", where m = n - 1 and r is an undefined
variable; e.g., when there are four peaks, n = 4 and m = 3, leading to r 3 +
3r2 + 3r + 1 from the expansion of (r + 1)'. The coefficients 1: 3: 3: 1 produce
the relative intensities. The Pascal triangle is a convenient device for remembering
the coefficients of the binomial expansion for nuclei with I = 1/2
1 1
1 2 1
1 3 3 1
1 4 6 4 1
1 5 10 10 5 1
1 6 15 20 15 6 1
The triangle is readily constructed, for the sum of any two elements in a row
equals the element between them in the row below.
When two groups of non-equivalent nuclei B and C split a third nucleus A,
the number of peaks in the A resonance is given by
nA = (2ESB + 1)(2S, + 1) (7-51)
This is equivalent to what was done in the discussion of the methylene resonance
of ethanol. Each of the four peaks from spin coupling by CH 3 was further
split into a doublet from coupling to OH, leading to a total of eight peaks.
When a nucleus with I > 1 is coupled to the observed nucleus, the number of
lines is still given by equation (7-50); however, the intensities are no longer given
by the binomial expansion. For instance, in Fig. 7-26(B), the hydrogen signal is
split into three lines by the deuterium, for which I = 1. The intensities of the three
lines are all equal and not in the ratio 1: 2: 1, since there is near equal probability
that the deuterium will have + 1, 0, and -1 magnetic quantum numbers. The
same situation occurs with splitting from "N.
For both chemical shift and spin-spin coupling applications, one must
recognize when nuclei in a molecule are non-equivalent. This will be the subject
of the next section.
7-15 DISCOVERING NON-EQUIVALENT PROTONS
In some molecules, a certain proton appears to be non-equivalent to others in

a particular rotamer but becomes equivalent when rapid rotation occurs; e.g.,
consider the non-equivalence of the three methyl protons in a staggered con-
figuration of CH 3 CHCl2 . An interesting problem in which the nonequivalence
is not removed even with rapid rotation is observed in the chemical shift of
substituted ethanes of the type:(18)
A X
\ /
B-C-C-H
/ \
D H
7-16 Effect of the Number and Nature of the Bonds on Spin-Spin Coupling 247
The two protons are not equivalent from symmetry arguments, and no
conformation can be found in which the two protons can be interchanged by a
symmetry operation. This intrinsic asymmetry will result in different chemical
shifts which, if large enough, will be observable. This asymmetry can be seen if
you label the two hydrogens and make a table indicating which atoms are
staggered on each side of the two hydrogens for all possible staggered rotamers
of (ABD)C-C(H2 X).
Early interpretations of the nonequivalence in the nmr assigned the different
shifts to different rotamer populations. Even in the absence of population
differences the protons are intrinsically non-equivalent in all rotamers.
The criterion to be applied in recognizing this phenomenon is the absence, in
any conformer that can be drawn, of a symmetry operation that interchanges the
two protons. In CH 3 CH 2 Cl, all the methyl protons are interconvertible in the
various rotamers that can be drawn, as are the methylene protons. Another
interesting example( 19 ) of non-equivalent protons involves the methylene protons
of (CH 3CH 2 O) 2 SO. The two protons of a given methylene group are not
stereochemically equivalent because of the lack of symmetry of the sulfur atom
with respect to rotation about the S-O-C bonds. One of the possible rotamers H 0
is depicted in Fig. 7-28. The small dot in the center represents the sulfur with CH 3
lines connecting the oxygen, the lone pair, and the ethoxyl groups. The large
circle represents the methylene carbon with two hydrogens, an oxygen, and a 0
H
methyl group attached. The molecule is so oriented that we are looking along a OC2 H,
line from sulfur to carbon. The non-equivalence of the two methylene hydrogens
of a given CH 2 group is seen in this rotamer, and they cannot be inter- FIGURE 7-28
changed by a symmetry operation in any other rotamer that can be drawn. Two Non-equivalence of the two
non-equivalent nuclei, which cannot be interchanged by any symmetry element protons of a given methylene
in one of the rotamers of
the molecule possesses, are called diastereotopic nuclei.
(CH3CH2 O) 2SO.
7-16 EFFECT OF THE NUMBER AND NATURE

OF THE BONDS ON SPIN-SPIN COUPLING
The nature of the spectra of complex molecules will depend on the number of
bonds through which spin-spin coupling can be transmitted. For proton-proton
coupling in saturated molecules of the light elements, the magnitude of J falls
off rapidly as the number of bonds between the two nuclei increases and, as
mentioned earlier, usually is negligible for coupling of nuclei separated by more
than three bonds. Long-range coupling (coupling over more than three bonds)
is often observed in unsaturated molecules. A number of examples are contained
and discussed in the literature.(2 0 ) In unsaturated molecules, the effects of nuclear
spin are transmitted from the C-H a bond by coupling of the resulting electron
spin on carbon from the sigma bond with the 7Eelectrons. This it electron spin
polarization is easily spread over the whole molecule because of the extensive
delocalization of the it electrons. Spin polarization at an atom in the it-system
can couple back into the a-system, in the same way that it coupled into the
it-system initially. In this way, the effect of a nuclear spin is transmitted several
atoms away from the splitting nucleus. This problem has been treated quantitat-
ively.?(21)
When spin-spin coupling involves an atom other than hydrogen, long-range
coupling can occur by through-space coupling. This interaction is again one of
spin polarization, but it occurs through non-bonding pairs of electrons, and the
coupling is through space instead of through a or 7r bonds. In the compound
CF 3 CF 2 SF5 the coupling constant for the fluorines of the CF 2 group and the
trans fluorine of the SF, group is about 5 Hz, much smaller than that for the
JP II I JP F
CF 2 group coupling to the cis fluorines (- 16 Hz), because of the reported
contribution from through-space coupling in the latter case.122 a This effect has
JP-H been invoked rather loosely, and it is suggested that an abnormal temperature
dependence be demonstrated before such an explanation is accepted.(22b)
In order to demonstrate one further point, it is interesting to consider the
phosphorus nuclear magnetic resonance spectrum of the compound HPF 2 . The
"stick spectrum" in Fig. 7-29 represents the splitting of the phosphorus signal
by the two fluorine atoms followed by the smaller 'JP-H splitting. The resulting
spectrum is indicated.
If instead of 'JPF > 1JPH we have 1 JP-H > 1JP-F, the spectrum in Fig.
7-30 arises. The fact that 'JPH is larger than 'JP F, is illustrated in the "stick
spectrum." The spectrum is expected to be one of those shown in Figs. 7-29 or
FIGURE 7-29 P nmr 7-30, depending upon the magnitude of 'JP-F VS- 'JP-H. If the (P-F coupling
spectrum expected for HPf2 constant is greater than the P-H coupling constant, Fig. 7-29 will result,
if JP-F> J-H (A) Stick whereas Fig. 7-30 results if 1JP-H is much greater than JP--F. If the two coupling
interpretation. (B) Spectrum. constants were similar, the spectrum would be a complex pattern intermediate
between those shown. In most compounds studied, JP--F > P-H (~ 1500 Hz
vs. ~200 Hz), and as expected, the spectrum in Fig. 7-29 is found experimentally.
If two nuclei splitting another group in a molecule are magnetically non-
equivalent, the spectrum will be very much different from that in which there is
splitting by two like nuclei. Four lines of equal intensity will result from splitting
by two non-equivalent nuclei with I = 1/2, and three lines with an intensity ratio
1:2:1 will be observed for splitting by two equivalent nuclei. Consequently, we
must determine what constitutes magnetically non-equivalent nuclei. Non-equiv-
alence may arise because of differences in the chemical shifts of the two splitting
nuclei or because of differences in the J values of the two splitting nuclei with
the nucleus being split. Equivalent nuclei have identical chemical shifts and
coupling constants to all other nuclei. Isochronous nuclei have the same chemical
shift and different couplings. In the molecule H 2C=CF 2 , where the two hydro-
gens and the two fluorines are isochronous and non-equivalent. Each proton
sees two non-equivalent fluorines (one cis and one trans) which have identical
chemical shifts but different JH-, coupling constants. The simple 1: 2:1 triplet
expected when both protons and both fluorines are equivalent is not observed
in either the fluorine or the proton nmr because of the non-equivalence that
JP- F
P-H
JP-H-
JP-F
(A) (B)
31
FIGURE 7-30 P nmr spectrum expected for HPF 2 if JP-H > P-F. (A) Stick
interpretation. (B) Spectrum.
exists in the J values. We shall consider this effect in more detail in the section
on second order effects in the next chapter.
Another effect that gives rise to spectra other than those predicted by
equation (7-50) is nuclear quadrupole relaxation. Often splittings do not occur
because the quadrupolar nucleus to which the element being investigated is
attached undergoes rapid relaxation, which causes a rapid change in the spin
state of the quadrupolar nucleus. Only the average spin state is detected; in some
instances this relaxing quadrupolar nucleus gives rise to very broad lines in the
nmr spectrum of the I = /2 nucleus bonded to it, and sometimes a proton
resonance absorption is broadened by this effect to such an extent that the signal
cannot be distinguished from the background. We shall discuss this in more
detail shortly in Section 7-25.
7-17 SCALAR SPIN-SPIN COUPLING MECHANISMS
In the preceding section the splitting in the spectrum of the HD molecule by the
magnetic moment of the attached nucleus was discussed. We shall be concerned
here with the mechanism whereby information regarding the spin of the nucleus
causing the splitting is transmitted to the nucleus whose resonance is split.
Consider a nucleus A split by the nucleus B (which has a spin of '/2) in the
molecule AB. The nuclear spin of B is transmitted to A by polarization of the
bonding electrons. Various processes cause polarization, and they constitute the
mechanisms for spin-spin splitting of hydrogen in solution. They also contribute
to the magnitude of the scalar spin-spin coupling constant, J, in various
molecules. Contributions to J are transmitted via the electron density in the
molecule and consequently are not averaged to zero as the molecule tumbles.
Three contributions will be considered:
1. Spin-orbital effects
2. Dipolar coupling, indirect, or through-space coupling
3. Fermi contact coupling
It is worth emphasizing that all three effects are transmitted via the electron density
in the molecule.
Spin-orbital effects involve the perturbation that the nuclear spin moment
makes on the orbital magnetic moments of the electrons around the nucleus.
For an I = /2 nucleus which we label B, for example, the magnetic field of the
nuclear dipole interacts differently for m, = + 1/2 than for m, = -'/2, with the
orbital magnetic moment of an appropriate electron around B causing a change
in the magnetic field from the orbital contribution of these electrons. The new
field from the electrons at B produces a field at the nucleus being split, A, that
depends upon whether the nuclear moment at B is +1 2 or - /2. Consequently, a
splitting of the nmr resonance of A results. The Hamiltonian for the interaction
on B that is felt at A is
gNN e L3 (7 52)
r
where L is the electron orbital angular momentum operator, f is the nuclear spin
moment operator, and r is the distance from A to B.
Dipolar coupling, often referred to as indirect coupling, is analogous to the

classical dipolar interaction of two bar magnets. Since the classical situation is
simpler, we shall treat it first. It is essential that you obtain a complete
understanding of this interaction, for we shall encounter many phenomena in
which this type of interaction is important in the remainder of this text. The
classical interaction energy, E, between two magnetic moments (which we shall
label i, and AN) is given by:
E = R3 (7-53)
r 3r5
where i is a radius vector from 9, to AN and r is the distance between the two
moments.
The indirect dipolar coupling mechanism corresponds to a polarization of
the paired electron density in a molecule by the nuclear moment. The polarization
of this electron density depends on whether m1 = + 1/2 or - /2, and the modified
electron moment is felt through space by the second nucleus. Replacing A, with
-g#$ for the electron magnetic moment and AN with gNJ3 NI for the nuclear
magnetic moment gives the dipolar interaction Hamiltonian:
I- flgNI3N f- 3(I-4)($4)(
The interaction between the electron spin moment and the nuclear moment
polarizes the spin in the parts of the molecule near the splitting nucleus B. When
this effect is averaged over the entire wave function, the field at B is modified
and this modified field of the electron density acts directly through space on the
nucleus A, which is being split. The direction of the effect depends upon the m,
value of the B nucleus, and thus this effect makes a contribution to J at the A
nucleus.
The important point is that this indirect dipolar coupling of the nucleus B
and the paired electron density, which modifies the electron moment felt at A,
is not averaged to zero by rotation of the molecule. This is illustrated in Fig.
7-31, where an m, = +1/2 value is illustrated at B. Three different orientations
of the molecule are shown. The lines of flux from the moment on B are shown
affecting the moment at the electron. The lines of flux associated with the change
in the moment at the electron from B are shown at nucleus A. For the three
orientations shown, the direction at A is the same.
/ /A, '
A e'- BB e- A
HO H0 HO-
FIGURE 7-31 The indirect dipolar coupling of the nuclear moment B to A for
different orientations of the molecule, AB.
The Fermi contact term is the final coupling mechanism we shall consider
for molecules rapidly rotating in solution. This mechanism involves a direct
interaction of the nuclear spin moment with the electron spin moment such that
there is increased probability that the electron near B will have spin that is
H {j HA
antiparallel to the nuclear spin. Since the electron pair in the bond have their FIGURE 7-32 The
spins paired, a slight increase in the probability of finding an electron of one spin antiparallel alignment of the
near B will result in there being an increased probability of finding an electron nuclear moment (solid
arrow) and the electron spin
of opposite spin near A, as shown in Fig. 7-32. Thus, A receives information
moment (hollow arrow).
about the spin of nucleus B; since the effect at A is in opposite directions when
m, equals +'/2 or - 1/2, this interaction contributes to the magnitude of J. The
effect is a direct interaction of the nuclear spin and the electron spin moment,
af -$9,where a is the coupling constant.
If we consider the spin of nucleus B to be quantized along the z-axis for
simplicity of presentation, we can describe this effect in more detail. The contact
Hamiltonian becomes:
A 38
3
9NN Bz (7-55)
where $zB has contributions from both electrons I and 2 at B given by
SzB S 1
9(r, - rB) + gz 2 b(r 2 - r.) (7-56)
Here rB is the radius of the nucleus and r1 is the distance from nucleus B to the
electron. The symbol 6(r, - rB) is the Dirac delta function and has a value of
zero unless electron I is at nucleus B. Thus, if p-orbitals were used to bond two
atoms together, there would be nodes at the nuclei and this term would be zero.
Accordingly, in molecules where the Fermi contact term dominates, e.g., J3CH,
correlations of J with the amount of s character in the bond have been reported.
The mechanism for uncoupling the spins in the u-bond involves field-induced
mixing of ground and excited states. Perturbation theory produces:
-8 gpgN N 2 _ Ain><nlg,,0> + K0Zn>KnISzA10> (7-57)

3 n En- E(
In the H2 molecule, JH-H (estimated from JH-D) has an experimental value

of 280 Hz. About 200 Hz comes from the Fermi contact term, 20 from the dipolar
contribution, and 3 from the nuclear spin-orbital effect. In general, the Fermi
contact contribution dominates most coupling constants involving hydrogen,
e.g., IJ1C-H, JSn-H, etc.
2
It can also be appreciated that the value of J will simply depend upon the
energy difference of the two different kinds of molecules containing B in different
spin states. This energy difference will be independent of-the field strength. The
field strength independence of J provides a criterion for determining whether
two peaks in a spectrum are the result of two non-equivalent protons or
spin-spin coupling. The peak separation in the spectrum will be different when
60-MHz and 100-MHz probes are employed if the two peaks are due to
non-equivalent protons, but the separation will be the same at different fre-
quencies if the two peaks arise from spin-spin splitting.
7-18 APPLICATIONS OF SPIN-SPIN COUPLING

TO STRUCTURE DETERMINATION
There have been many applications of spin-spin coupling to the determination

of structures. For example, if the spectrum of a sample contains the very
characteristic fine structure of the CH 2 and CH 3 resonances of an ethyl group,
this is a good indication of the presence of this group in the molecule. Other
applications involve variations in the magnitude of J in different types of
compounds. For two non-equivalent protons on the same sp2 carbon, e.g.,
ClBrC=CH 2, proton-proton spin coupling constants, JH-H, of I to 3 Hz are
observed. Coupling constants for non-equivalent trans ethylenic protons have
values in the range of 17 to 18 Hz, whereas cis protons give rise to coupling
constants of 8 to 11 Hz. These differences aid in determining the structures of
isomers.
o 0 The characteristic chemical shift of hydrogens attached to phosphorus
occurs in a limited range, and the peaks have fine structure corresponding to
H H 'JP-H (phosphorus I = '/2). The phosphorus resonance 23 ) of HPO(OH) 2 and
0 OH
H OH H 2 PO(OH) is a doublet in the former and a triplet in the latter compound,
H
supporting the structures in Fig. 7-33. The coupling of the hydroxyl protons
(A) (B)
with phosphorus is either too small to be .resolved, or it is not observed
FIGURE 7-33 Structures of because of a fast proton exchange reaction, a phenomenon to be discussed
(A) phosphorous and (B) shortly. Similar results obtained from phosphorus nmr establish the structures of
hypophosphorous acids. FPO(OH)2 and F 2PO(OH) as containing, respectively, one and two fluorines
attached to phosphorus. The "P resonance in P4 S 3 consists of two peaks with
intensity ratios of three to one.( 4 ) The more instense peak is a doublet and the
less intense is a quadruplet. Since I = 0 for 3 2 S, both the spin-spin splitting and
the relative intensities of the peaks indicate three equivalent phosphorus atoms
and a unique one. It is concluded that P4 S 3 has the structure in Fig. 7-34.
P The fluorine resonance in BrF, consists of two peaks with an intensity ratio
\ ~S'
of four to one. The intense line is a doublet and the weak line is a quintet
(1:4:6:4:1). Quadrupole relaxation eliminates splitting by bromine (I for
79
Br = "Br = 3/2). This indicates that the molecule is a symmetrical tetragonal
pyramid.
P -- P Solutions of equimolar quantities of TiF 6 2 and TiF 4 in ethanol give
fluorine nmr spectra( 2 s> consisting of two peaks with the intensity ratio 4: 1 (1 = 0
PL for 4 8Ti). The low intensity peak is a quintuplet and the more intense peak a
doublet. The structure [TiF,(HOC2H)] - containing octahedrally coordinated
FIGURE 7-34 The structure titanium was proposed.
of P4S3. The factors that determine the magnitude of the coupling constant are not
well understood for most systems. It has been shown that in the Hamiltonian
describing the interaction between the 3C nucleus and a directly bonded proton,
the Fermi contact term is the dominant one. Qualitatively, this term is a measure
of the probability of the bonding pair of electrons existing at both nuclei. The
need for this can be appreciated from the earlier discussion of the Fermi contact
coupling mechanism. The greater the electron density at both nuclei, the greater
the interaction of the nuclear moments with the bonding electrons and hence
with each other through electron spin polarization. Since an electron in an s
orbital has a finite probability at the nucleus and p, d, and higher orbitals have
nodes (zero probability) at the nucleus, the Fermi contact term will be a measure
of the s character of the bond at the two nuclei.
7-18 Applications of Spin-Spin Coupling to Structure Determination 253
Since the s orbital of hydrogen accommodates all of the proton electron
density, the magnitude of IJ1c-H for directly bonded carbon and hydrogen will
depend upon the fraction of s character, p, in the carbon hybrid orbital bonding
the hydrogen. The following equation permits calculation of p from 'J,,C-H data.
50 0
'JC-H3_(Hz) = PC-H (7-58)
Since p = 0.33 for an sp 2 hybrid and 0.25 for an sp3 hybrid, it can be seen that
should be a sensitive measure of carbon hybridization. The use of IJ3CH
HJc-H
to measure the hybridization of carbon in a C-H bond was studied in detail(2 ">
and is supported by valence bond calculations!") As expected from equation
(7-58) the 3 C-H coupling constant for a hydrogen in a saturated hydrocarbon
is about 125 Hz, that for the ethylenic hydrogen of a hydrocarbon is around
160 Hz, and that for an acetylenic hydrogen is around 250 Hz, corresponding to
p's of 0.25, 0.33, and 0.50, respectively.
It has been shown that a linear relationship exists between IJIC-H and the
proton shift, 6, for a series of methyl derivatives in which the contribution to r
from magnetic anisotropy is approximately constant,(12 8 , 291) or varies with the
inductive properties of the group attached to methyl. Those compounds that
deviate from this line have appreciable contributions to 6 from anisotropy. The
following explanation is offered for the existence of this relationship. Isovalent
hybridization arguments 3 ) indicate that as X becomes more electron with-
drawing in the compounds CH 3X, more p character is employed in the C-X
orbital, and there is a corresponding increase in the s character in the C-H
orbitals. Carbon becomes more electron withdrawing toward hydrogen as the
hybridization changes from sp 3 to sp 2 to sp.(31 ) If there are no anisotropic
contribution, a correlation is expected between carbon hybridization and (. Since
carbon hybridization, is in turn related to 'JC-H, a correlation is expected
between IJ13C-H and 6 for compounds in which the contributions to ( from
magnetic anisotropy are constant (or vary linearly with the electron-withdrawing
properties of group X). In view of the difficulty of assessing the existence of even
large anisotropic contributions to the chemical shift (see section on chemical
shifts which are greatly influenced by anisotropy), this relationship is of consider-
able utility. Certainly the chemical shifts of compounds that do not fall on the
line cannot be interpreted in terms of inductive arguments, i.e., a local diamagne-
tic effect.
The constancy of the 1 IJC-H coupling constants of the alkyl halides (149,
150, 152, and 152 Hz respectively for F, Cl, Br, and I) is interesting to note. Even
though there is a pronounced change in the electron withdrawal by F, Cl, Br, and
I, 'J C-H does not change(2 " as would be predicted from the isovalent hybridiza-
tion argument. In this series, there is an effect on the carbon hybridization that
is opposite in direction to electron withdrawal by the halogen. These two
effects cause the hybridization and electron-withdrawing properties of carbon
toward hydrogen to remain constant. This other effect is based on orbital overlap,
and one should think in terms of the s orbital dependence of the bond strength
for the three hydrogens and X. As X becomes larger, the orbital used in sigma
bonding becomes more diffuse and the overlap integral becomes smaller. Con-
sequently, the energy of the C-X bond becomes less sensitive to the amount of
s-orbital used to bond X, and the total energy of the system is lowered by using
more of the s orbital in the bonds to hydrogen. Thus, as the principal quantum
number of the atom bonded to carbon increases, we expect to see, from this latter
effect, an increase in IJ3 c-H-
It has been shown that the carbon hybridization in Si(CH 3 )4 , Ge(CH 3)4 ,
Sn(CH 3)4 , and Pb(CH 3 ) changes in such a way as to introduce more s character
into the C-H bond as the atomic number of the central element increases. An
explanation for this change in terms of changing bond energies is proposed, and
it is conclusively demonstrated that chemical shift data for these compounds
cannot be correlated with electronegativities of the central element,<29 >because
large anisotropic contributions to the proton chemical shift exist in Ge(CH 3)4 ,
Sn(CH 3)4, and Pb(CH 3 )4 .
The 3 C-H coupling constant in the carbonium ions (CH 3)2 CH+ and
(C6H,) 2 CH' are 168 and 164 Hz, respectively, compared to values of 123 and
126 Hz for propane and (CH 5 )2CH 2 . This is consistent with a planar sp 2
hybridized carbon.
The magnitude of the coupling constant between two hydrogens on adjacent
carbons, 3 JH-H, is a function of the dihedral angle between them(1 2 34> as shown
in Fig. 7-35. The dihedral angle 0 is shown at the top of Fig. 7-35, where the
C-C bond is being viewed end-on with a dot representing the front carbon and
the circle representing the back one. The function is of the form:
3JH-H = A + B cos 0 + C cos 20 (7-59)

For hydrocarbons, A is 7 Hz, B is 1 Hz, and C is 5 Hz. The end values change
in different systems, but the shape of the curve is the same. Difficulties are
encountered with electron-withdrawing substituents and with systems in which
there are changes in the C-C bond order.
The 'J"c-H values can be of use in assigning peaks in a spectrum.13 51 For
example, the proton chemical shifts of the two methyl resonances in
CH 3C(O)SCH 3 are very similar, as shown in Fig. 7-36, and impossible to assign
directly to methyl groups in the compound. However, by virtue of the fact that
sulfur is a larger atom than carbon and has a comparable electronegativity, the
6 H
13
11
3
J
0 7r
2
6
FIGURE 7-35 variation in 3 JH-H with the dihedral angle 0
shown at the top.
7-18 Applications of Spin-Spin Coupling to Structure Determination 255
FIGURE 7-36 Proton nmr

spectrum of S-methyl
thioacetate at high spectrum
amplitude showing satellites
arising from "C. fFrom R. L.
Middaugh and R. S. Drago, J.
Amer. Chem. Soc., 85, 2575
(1963).]
=131 C.P.S.
'JC-H values for methyl groups attached to sulfur [J = 138 to 140 Hz in

CH 3 SH, (CH 3 ) 2SO, (CH 3) 2 SO2 ] are larger than those for methyl groups attached
to carbonyls (J = 125 to 130 Hz). Since the proton resonance on a "C will be
in the center of the "C satellites, the less shielded peak is unequivocally
assigned( 35 ) to the acetyl methyl.
It has been proposed136 371) that a relationship exists between Jsn-H and the
hybridization of tin in the tin-carbon bonds of compounds of the type
CH 34 -nSnXn. This relationship and results from other physical methods were
employed to establish the existence and structure of five-coordinate tin addition
compounds (CH 3)3 SnCl -B,( 7 ) where B is a Lewis base [e.g., (CH 3)2 SO or
CH 3C(O)N(CH 3)2]. The 2Jsn-H values suggest a trigonal bipyramidal geometry
in which the tin employs essentially sp 2 -orbitals in bonding to carbon, and,
consequently, a three-center molecular orbital using a tin p or p-d orbital in
bonding to the Lewis base and chlorine. Similar results were obtained for 2 JPb-H
in the analogous lead compounds.<38 )
Investigation( 39) of a whole series of addition compounds between
(CH 3)3SnCl and various bases indicated that the 2Jsn-H coupling constant
changed in direct proportion to -AH of adduct formation with the base. This
was interpreted to indicate that as the tin-base bond became stronger, the hybrids
used to bond the methyl groups approached sp 2 from the - sp3 hybrids used in
(CH 3)3SnCl. The relation found was:
2j] 19sa-H = 216 Psn (7-60)
The 2jSn-H values for Sn(CH 3 )4 , Sn(CH 3)3Cl, and Sn(CH 3 )3Cl 2 were 54, 57.6
(in keeping with the isovalent hybridization prediction), and 72 [which when
substituted into equation (7-60) gives Psn = 0.33]. The coupling constants for the
adducts ranged from 64.2 to 71.6; the weakest base studied was CH 3CN, and
the strongest was [(CH 3)2N] 3PO. The plot of -AH vs. 2
JSn-H was extrapolated
to the value for free trimethyltin chloride at -AH = 0.
Coupling constants of equivalent hydrogens cannot be determined directly.
Deuteration of a molecule has only a slight effect on the molecular wave function,
so this technique can be used to gain information about the coupling of
equivalent protons. For example, the H-H coupling in H 2 cannot be directly
measured, but that in HD is shown to be 45.3 Hz. Since the effect of isotopic
substitution on the magnitude of J is proportional to the magnetogyric ratio,
we have
JHH YH
- = 6.515 (7-61)
JHD YD
From this, we calculate 'JH-H = 295.1 Hz.

There have been several very interesting applications of phosphorus nmr in
the determination of structures of complexes of phosphorus ligands. 0 The nmr
spectrum and its interpretation are illustrated in Fig. 7-37 for the complex
Rh((p 3 P) 3C 3 [I for 1 3(Rh) = 1/21. Two isomers are possible, facial and mer-
idional. All of the phosphorus ligands are equivalent in the facial isomer, so the
spectrum in Fig. 7-37 substantiates that the complex studied was meridional. In
this isomer two phosphorus atoms are trans to one another (labeled Pb), and one
is trans to a chlorine (labeled Pa). The splittings are interpreted in the stick
diagram at the top of Fig. 7-37.
The magnitude of 2 J,_, often provides interesting information about the
stereochemistry of complexes. The magnitude of this coupling is usually much
larger when two phosphorus atoms are trans to one another than when they are
Pa Pb
JRhPa RhPb
JPRhPb a hPb
FIGURE 7-37 The 31p

nmr spectrum of
Rh((P3 P) 3 Cl 3 .
7-19 Effect of Fast Chemical Reactions on the Spectrum 257
cis. In cis-PtCl 2 (bu3 P)[(C 6 H 5 0) 3 P], the value of 2 J,_, is 20 Hz, while a value of
758 Hz is found in trans-PdI 2(bu 3P)[(C 6H10) 3P] and 565 Hz is found in cis-
PdI2[(CH 3)3P][(C 2 H,) 3P). Some exceptions to this rule are found in Cr', MoO,
WO, and Mn(I) compounds. For example, in trans-W(CO) 4 [(C6 H5 )3 P](bu 3 P),
2
Jpp is found to be 65 Hz.
A very interesting result is obtained in the proton or carbon-13 nmr spectra
of phosphorus complexes. In the proton nmr of trans-PdI 2[P(CH 3)3 2, we might
expect to find a doublet methyl resonance with perhaps a small splitting of each
peak from the second phosphorus. Although the two phosphorus atoms are
chemically equivalent (isochronous)they are not magnetically equivalent. Any given
methyl group would experience two different phosphorus couplings, 2 J_H and
2 2
'JP-H . When the J,_, is much larger than j P-H, as it is in the above complex,
the proton nmr spectrum observed is a 1: 2:1 triplet; i.e., the two phosphorus
nuclei behave as though they were two equivalent nuclei splitting the proton
resonance. The two phosphorus nuclei are said to be virtually coupled.(4 ) In a
similar cis complex where 2 j,_, < 2Jp-, , a doublet is obtained. In view of the
very common, large 2J,_, for trans phosphorus ligands and the small 2
values for cis phosphines, this phenomenon can be used to distinguish these two
kinds of isomers. The use of 3 C nmr in this type of application14 2) has been
criticized.(4 1) Triplets occur in the 3 C nmr with much smaller values of J,_p,
so it is more probable that cis complexes will show virtual coupling. The reasons
for virtual coupling have been discussed in the literature and will be considered
in more detail in the section on second order effects. Virtual coupling is an
attempt to treat second-order spectra in a first-order way. The second-order
analysis should be done.
7-19 EFFECT OF FAST CHEMICAL REACTIONS Factors Influencing

ON THE SPECTRUM the Appearance
If one examines the high resolution spectrum of ethanol in an acidified solution, of the NMR
the result illustrated in Fig. 7-38 is obtained, in contrast to the spectrum shown Spectrum
in Fig. 7-24. The difference is that the spin-spin splitting from the hydroxyl
proton has disappeared. Acid catalyzes a very rapid exchange of the hydroxyl
proton. In the time it takes for a methylene proton to undergo resonance, many
different hydrogen nuclei have been attached to the oxygen. As a result, the
-OH
-CH3
FIGURE 7-38 High resolution
I ICH2 nmr spectrum of a sample of
19i I ethanol (acidified).
Inc. field strength-

methylene proton experiences a field averaged to zero from the O-H nuclear
moment, and the 3JHCOH coupling disappears. In a similar fashion, the hydroxyl
proton is attached to many different ethanol molecules, averaging to zero the
field it experiences from CH 2 protons, and only a single resonance is observed.
A very dramatic illustration of this effect is the spectrum of a solution of
aqueous ammonia in which one does not see separate N-H and O-H protons,
but only a single exchange-averaged line. When exchange is rapid, the chemical
shift of this exchange-averaged line is found to be a mole-fraction-weighted
average of the shifts of the different types of protons being exchanged:
6AvG = NNH 3 6NH3 + NH,0 (762)
It is important to emphasize that NNH3 is not the mole fraction of ammonia, but
the mole fraction of N-H protons, i.e., NNH, = 3[NH,]/(3[NH 3 ] + 2[H 2 0]).
The "F spectra of solutions of TiF 4 in donor solvents taken at - 30' C consist
of two triplets of equal intensity. 4 ' Six-coordinate complexes form by coordinat-
ing two solvent molecules, and the spectrum obtained is that expected for the
cis structure. This structure contains two sets of non-equivalent fluorine atoms,
with two equivalent fluorines in each set. At 00 C only a single fluorine peak is
obtained. It is proposed that a rapid dissociation reaction occurs at 0 C, making
all fluorines equivalent;
TiF 4 -2B ' TiF 4 B + B
At -30' C this reaction is slowed down so that the non-equivalence can be

detected by nmr. Internal rearrangements and ionic exchange mechanisms are
also possible.
This example illustrates one of the possible pitfalls in structure determination
using nmr spectroscopy. If only the high temperature (0 C) spectrum had been
investigated or if rapid exchange occurred at - 300 C, it could have been
incorrectly assumed that the adduct had the trans structure on the basis of the
single nmr peak. If the actual structure were trans, only a singlet fluorine
resonance would be detected at all temperatures and it would have been difficult
to draw any structural conclusion because of the possibility that rapid exchange
might be occurring at both temperatures. Even in the present case, the possibility
exists, on the basis of these data alone, that the cis isomer is the structure at
-300 C and that the trans isomer predominates at 0' C.
The fluorine nmr spectra of a large number of compounds of general formula
R5 _,PF, (where R is a hydrocarbon, fluorocarbon, or halide other than fluorine)
have been reported.(44 1 The number of peaks in the spectrum and the magnitude
of the coupling constants are employed to deduce structures. For a series of
compounds of the type R 2 PF 3 , a trigonal bipyramidal structure is proposed, and
it is found that the JP.F values for axial fluorines are ~170 Hz less than those
for equatorial ones. The most electronegative groups are found in the axial
positions. The spectra and coupling constants, obtained on these compounds at
low temperature, indicate that the two methyl groups in (CH 3)2 PF 3 are equato-
rial and the two trifuoromethyl groups in (CF 3 )2PF 3 are axial. At room tempera-
ture, rapid intramolecular exchange occurs and the effect of this exchange is to
average the coupling constants.
7-20 QUANTUM MECHANICAL DESCRIPTION OF

COUPLING
When the magnitude of the separation between two peaks in the nmr, expressed
in Hz, is of the same order of magnitude as the coupling constant, so-called
second-order spectra result. When this occurs, the peaks in the resulting spectrum
cannot be assigned by inspection as we have done before. This is illustrated in
Fig. 7-41, where the spectra of CIF 3 , obtained by using two different probes,
demonstrate these complications.
With a higher-frequency probe (40 MHz), J < A and the spectrum in Fig.
7-39(A) is obtained. The molecule CIF 3 has two long Cl-F bonds and one short
one, giving rise to non-equivalent fluorines. The spectrum obtained at 40 MHz
is that expected for non-equivalent fluorines splitting each other. As expected,
the triplet is half the intensity of the doublet. Using a lower-frequency probe
(10 MHz and the corresponding lower magnetic field), the difference in A for the
non-equivalent fluorines is of the order of magnitude of JF-F (recall that A is
field dependent) and the complex spectrum in Fig. 7-39(B) is obtained. Complex
patterns of this sort result whenever the coupling constants between non-
equivalent nuclei are of the order of magnitude of the chemical shift.
(A) FIGURE 7-39 Fluorine

nmr spectra of CiF 3. (A) at
40 MHz and (B) at 10 MHz.
(B)
When second-order spectra result, other strategies than going to a higher

field can be used to obtain the chemical shifts and coupling constants. Our
discussion will involve protons, but the treatment is perfectly general for
any spin 1/2 nucleus.
The energy for a proton, A, surrounded by paired electron density in a
magnetic field is given by:
E = -gNflNHO(l - CA)mI
or the frequency of the transition, v, is given by:
v = v0(1 - UA)IAmI (7-63)
where vo is the resonance frequency for a bare proton and CA the shielding
constant.
Next consider a molecule containing two different hydrogens that are

involved in spin-spin coupling. We shall focus attention only on the two
hydrogens and ignore the rest of the molecule. When the chemical shift difference
of the two hydrogens is very large compared to their J, we shall label such a
system AX. The energy of a system of Xj protons whose shift differences are
larger than J is given by equation (7-64).
E = - h Y vo(1 - aj)m + h Y Jkmlmk (7-64)

j j<k
The first summation on the right gives a different chemical shift term for each
of the j different types of nuclei in the molecule. The second summation is taken
only for pairs of nuclei in which j < k, to insure that each pair is considered only
once. It is assumed in writing this equation that J is isotropic. The significance
of equation (7-64) is illustrated in Fig. 7-40, where the energies of the various
nuclear configurations of an AX system in a magnetic field are illustrated for the
case in which JAX is zero and for the case in which it is finite.
FIGURE 7-40 Energies for [ 1 o + hvo _ _)

the AX system in a magnetic [1 2 a-)] hvo
field [(A) = '/2, /(X) = '21-
(A) JAx = 0. the wave function,
written in the order AX, is >- (oG Ox) hv0
written above the energy
level; the energy is written
-- hJ
below the level. (B) JAx # 0. z + (CA -X) hv 4
The J labeling indicates the LU AX
change in energy that occurs IrhJ ()(JhJ
from the dotted line, which -1+ 2 (oA + ox)] hvo
represents the energy when
JAX 0. JAX = 0 JAX 0
(A) (B)
In Fig. 7-40A, where JAx = 0, the energy levels for the xc2, ocf#, #a, and ##
nuclear spin configurations are shown. The A nuclear spin state is listed first and
that of the X nucleus second. The ii energy is the sum of
-h[v 0 (1 UA)( /2) + v0 0 - UX)('/2) - [ -- 1 + (CA + Cx)('/ 2 )]hvo
The energies of the other levels are calculated similarly by substituting the
appropriate m, values. The solid arrows indicate the transitions for the A type
of nucleus. One arrow corresponds to A nuclei bonded to X nuclei with ctspin
and the other to those on X nuclei with # spin. The section rule is Am, = 1, so
the A and X nuclear spins cannot change simultaneously. The dashed arrows
indicate the transitions of the X nucleus. By referring to equation (7-64) and Fig.
7-40, we see that the frequency difference between 22 and f#2 (v, - CAVo) is equal
to the difference between fl and ##/3; i.e., the two transitions of nucleus A are
degenerate. Thus without any coupling, a single peak would be observed for A.
By similar reasoning, we expect a single peak for X.
Where JAX is finite, and considerably smaller than the chemical shift
difference of A and X, the Jjk mimk terms modify the energies shown in Fig. 7-40A,
producing those shown in (B). Since the c2 energy is raised by ('/4) Jh and #Lx is
lowered by (1/4) Jh, the 2c -- # 2lotransition occurs at a frequency J/2 lower than
the corresponding transition in the JAX = 0 case. The 2# -- transition is at a
frequency J/2 higher than that of the JAX = 0 transition. Thus, the frequency
separation between the two peaks is equal to J in the JAX A 0 case.
Spin-spin coupling constants can be positive or negative. In the above
problem, a positive J was assumed. If J was taken as negative, the Lxotenergy
would be lowered and the #lcenergy would be raised. This would shift the peak X
A
for the 22 --* #Lxtransition to a frequency J/2 higher than in the JAX = 0 case, i I
whereas the other transition would occur at a frequency J/2 lower. (A greater
U--JAx--I I--JxaA
frequency corresponds to a lower field.) There would be no change in the 1 2 3 4
appearance of the spectrum, so there would be no way to determine the sign of (A)
J as illustrated by referring to Fig. 7-41. In Fig. 7-41(A), line I represents the
A X
chemical shift difference between A and X and line II shows the spin-spin splitting I I
of A by X when JAX is positive. If JAX were negative the result in Fig. 7-42(B)
JAX1 t-- JXAA 1
would be obtained. One cannot distinguish the two possibilities by examining 2 1 4 3
the spectrum obtained in the normal nmr experiment. (B)
Equation (7-64) is the so-called first-order solution of the chemical shift-
coupling constant problem. It applies only when the chemical shift difference, A, FIGURE 7-41 Effect of the
sign of Jx on the splitting of
of the two nuclei is large compared to their coupling constant, J (usually A - 5J).
an A-X system. Here 1
We now treat the more general problem using a system where A and J have refers to the = to #2a
comparable magnitudes. The label AB will be used to describe this case, where transition, and 2 refers to the
I(A)= /2 and I(B)= /2. We shall solve for the energies using: af#to ## transition.
O*R dr = 0*0 di
hf
With E/h being a frequency, the Hamiltonian is written in frequency units as:
H = - v0 (1 - aA)IZA - v0 0 -- B)I + AB A -TB (7-65)
Our basis set will be the nuclear spin functions p1 = |2co), 2 = |cc>, 93 = |c>,
and 94 =|#>. To obtain the four energies, we need to solve for all the matrix
elements in the secular determinant:
E
Kccc'iiflPi
h
(xx| Ahim
114 Icccc
cc13 ( l11fl aI
(o|lpp)l
ficcIIiIcccc> Ic>
(xp|AlIH K/IcIHI/Icc>- E
h
Kflifilifl/3
1ppi3/11pa> (ppifi>
The term - E/h appears on the diagonal because only the diagonal
[-(E/h)(cp~Ip.>] terms are non-zero; for example,
-(E/h)a)cctc~> = -E/h
while
-(E/h)(af#|> = 0
The solution of this secular determinant for the four energies is shown in
Appendix G. The four energies are given by:
E,
F = V0 [-1 + ('/2)aA + ('/2)Bl + (766a)
h 4
E,
_(1/4)J - C (7-66b)
h2
where C (1/ 2 )(j

2
+ A2 )'/2
E3
3= _(1/4)J + C (7-66c)
h
E,
=
-v[1 - (1/2)A - (' /2)B +-66d)
h 4
The corresponding wave functions are listed below.
0,1 = |01c> (7-67a)
0 2 = cos 01x# - sin OIIpa> (7-67b)
3= sin O6cxl> + cos 6|#> (7-67c)
04 = I## (7-67d)
We can summarize the results by constructing an energy diagram for this AB

system in Fig. 7-42, as was previously done in Fig. 7-40 for the AX system. In
Fig. 7-42(A), the situation with JAB = 0 (which is identical to Fig. 7-40(A)) is
presented as a starting place. The contribution to the total energy of the various
levels from JAB is shown in Fig. 7-42(B).
The arrows for the four transitions are indicated in Fig. 7-42. When A and
B have similar chemical shifts, all four transitions wifl appear in a narrow region
of the spectrum. The center of the spectrum is given by the average of the E, -+ E 2
and E 2 -- E4 transition energies or:
E - E,
('/2)(E2 - E1 + E4 - E2) = = vo[1 - (1/2)OA - ('/2)oB
7-21 Effects of the Relative Magnitudes of J and A on the Spectrum of an AB Molecule 263
-gE 4 =[1 (A B) 0
[1 2oA + GB)] O
E, = i + C*
2
(C>J)
+y(o, - o') vo - E2 = -j-C
+El = [-1 + j (oA B
(A) (B)
FIGURE 7-42 The second-order energies (written in frequency units) for an AB
spin system. (A) JaB = 0. (B) The second-order result. (*These are the total
energies, where the quantity C contains the shielding constant terms.)
The transition energies relative to this center can then be found using the energies
given above and illustrated in Fig. 7-42(B). The results are presented in Table
7-5 along with the intensities of the various transitions. The intensities are
dependent on fx (recall that we discussed the fact that H1 made the Ix matrix
elements non-zero) and are determined from the evaluation of integrals of the
form
I0jjfIx
+ IXBkLi>12 (7_68)
For example, the 1 - 3 (i.e., E, to E,) transition yields the result:
|(1|IxA + IxB cos 0 + x# sin 0>2 _ (1/4 )(cos 0 + sin 0)2

= ('/4)(1 + sin 20)
7-21 EFFECTS OF THE RELATIVE MAGNITUDES OF J

AND A ON THE SPECTRUM OF AN AB
MOLECULE
We are now in a position to use the results summarized in Table 7-5 to see what
influence the relative magnitudes of A and J have on the appearance of the
spectrum.
TABLE 7-5. Energies and Intensities for the

Transitions in an AB Molecule
Transition Separation from Center Relative Intensity
1 --+2 -(J/2) - C 1- sin 20

1 -3 - (J12) + C 1+ sin 20
2 - 4 (J/2) + C 1- sin 20
3- 4 (J/2) - C I+ sin 20
where sin 20 = J/2C and C = (1/2)(j
2
+ A2 )1 /2
A. J=OandAO
The 1 -+ 2 and 3 -+ 4 transitions are degenerate, as are 1 -- 3 and 2 - 4. The
spectrum consists of two peaks, one from A and one from B, as in the earlier
discussion of Fig. 7-40(A).
B. J# 0, A : 0, but J << A-an AX System

When A dominates C, we have sin 20 = J/A ~ 0. With the sin 20 equal to zero,
all four transitions have equal intensity. The energies become those shown in
Fig. 7-40(B). Relative to the center, we have 1 -- 2 at (-'/2)J - ('/ 2 )A, 1 - 3 at
(- 1/2)J + ('/ 2 )A, 2 - 4 at ('/2)J + ('/ 2 )A, and 3 - 4 at (1/2)J - ('/ 2 A) for a
spectrum that looks like that shown in Fig. 7-43.
Hinc
FIGURE 7-43 Spectrum of an

AX system. (Greater frequency
corresponds to lower field.)
2-+4 1-+3 3 -4 1- 2
C. AOandJ O,butJ zA
When J z A, sin 20 is appreciable and positive. Therefore, the 1 -- 3 and 3 -* 4
transitions will have equal intensities which are greater than those of 1 - 2 and
2 -4. The 1 -- 2 and 3 -> 4 transitions will differ by J, as will the 2 - 4 and
1 - 3; also, the 3 -> 4 and 2 -+ 4 will differ by 2C, as will the 1 -+ 3 and 1 -+ 2.
The spectrum will look like that in Fig. 7-44.
1 + J/2C 1 - J/2C
FIGURE 7-44 Spectrum of

an AB system. I--J -+ :- 2C I ) |
2 -+4 1 -+-- 3 3 ->.4 1 -+2
D. J # 0, A very small (A < J)

When CA and aB are almost equal, the value of sin 20 approaches unity and the
1 -> 2 and 2 -+ 4 transitions can be easily lost in the baseline noise. The spectrum
1 -3 3 -- 4 resembles that shown in Fig. 7-45.
FIGURE 7-45 Spectrum of an
AB molecule when UA and UB #0, but CA=
E. J7 B,so A = 0
are close.
This is the case of equivalent protons, and we are now in a position to see why
equivalent protons do not split each other. When A = 0, sin 20 equals unity and,
as can be seen from Table 7-5, the 1 -+ 2 and 2 -> 4 transitions have zero intensity.
The 1 - 3 and 3 -+ 4 transitions have intensity, but they have the same energy
and occur at the center of the spectrum; i.e., -('/2)J + C = ('/2)J - C = 0. This
is a manifestation of a general rule: magnetically equivalent nuclei do not split
each other.
7-22 More Complicated Second-order Systems 265
7-22 MORE COMPLICATED SECOND-ORDER
SYSTEMS
It is convenient to classify molecules according to the general type of A and J

systems to which they belong, for this suffices to define the Hamiltonian for the
coupling. The conventions that will be given are normally applied to spin 1/2
systems.
1. Each type of magnetic nucleus in the molecule is assigned a capital
letter of the Roman alphabet. All of one equivalent set of nuclei are given a single
letter, and a subscript is used to indicate the number of such nuclei in the set;
e.g., benzene is A 6 -
2. Roman letters are assigned to different nuclei in order of decreased
shielding; i.e., three different nuclei with similar shifts listed in order of decreased
shielding would be labeled A, B, and C. If the second set of nuclei has a very
different shift than the first (i.e., A > J), a letter far removed in the alphabet is
used (X, Y, Z) to label the low field peaks. Nuclei in between would be labeled
L, M, N, etc. For example, acetaldehyde, CH 3 CHO, is an A3X system. This
implies a first-order system. HF would be an AX system. The molecule HPFC 2
would be an AMX system. When one is not sure of the chemical shift difference,
nearby letters are used to be safe, for this implies that the system could be second
order.
3. The system of labeling must provide information about the magnetic, F(l)
as well as the chemical, non-equivalence of nuclei in the molecule. If two atoms (1) H Cl
are equivalent by symmetry, they have the same chemical shift, but they could
be involved in spin-spin coupling to other nuclei such that the J to one nucleus (2) H Cl
is different from the J to another; e.g., in CH 2 F 2 C 2 (Fig. 7-46) the two F(2)
hydrogens are isochronous, but two different HF couplings are needed for the
FIGURE 7-46 The molecule
full analysis. The H(1)-F(1) coupling constant is different from the H(1)-F(2)
1,2-dichloro, 3,6-fluoro
coupling constant. The protons and the fluorines are said to be chemically benzene; an AA'XX' system.
equivalent, but magnetically non-equivalent. All that is meant by this is that, in
the Hamiltonian for this system, two different HF couplings must be considered
even though the two hydrogens and two fluorines are related by symmetry. The
dichloro-difluoro-benzene isomer discussed above is thus referred to as an
AA'XX' system. The reader should now realize that in the non-planar (D2 )
molecule
H/H c)
13C=C="C
/\Hd
1'Ib)
the ab pair is magnetically non-equivalent to the cd pair because two different

coupling constants to the 3 C atoms are involved. The system will be labeled
A2A 2'XX'. The molecule ClF 3 discussed in Chapter 6 (consider only fluorines)
would be classified as an AB 2 molecule (at 10 MHz). The phosphorus nmr of
the tetrapolyphosphate anion, P4 0 136 , would yield an A2 B2 spectrum, and
P4 S3 is an A 3X system.
The energy levels of various kinds of coupled systems have been worked
out in detail( 8 -4 5, 4 6) in terms of the shift differences and coupling constants, and
are reported in the literature. To use these results, one classifies the molecule of
interest according to the scheme described above and looks up the analysis for
8
this type of system; e.g., one could find the energy levels for any AA'XX' system.( )
Computer programs are also available that find the best values of the J's and
A's that reproduce the chemical shifts and intensities of all the peaks in the
experimental spectrum. The interpretation of a relatively simple second-order
O 0(3-)
3
/ /
spectrum is illustrated for the P nmr spectrum of the anion H-P-P-0
0 0
indicated in Fig. 7-47. The actual spectrum, IV, is interpreted by generating it
in three stages. Consideration of the first stage, I, yields two lines from the two
non-equivalent phosphorus atoms, P(a) and P(b), and their separation is the
chemical shift difference. The second consideration, II, includes splitting by
hydrogen. Since the hydrogen is on P(a), JP--H >JPI-H . The four lines which
result are included in 1I. The third consideration, III, included P(a>-P(b) splitting
and accounts for the final spectrum for this ABX case. Two of the expected lines
are not detected in the final spectrum because they are too weak to be detected
and another pair fall so close together that they appear as a single peak (the
most intense peak). The analysis of the HP 2 0 - spectrum to yield the inter-
pretation contained in Fig. 7-47 was carried out(") with a computer analysis
that fitted the intensities and chemical shifts of the experimental spectrum.
The phenomenon of virtual coupling, which we discussed earlier, is a
magnetically non-equivalent system of, for example, the type XAA'X' where JAA'
is large.
0 3-
H 0
0 0
J HP~~K) HP
- -
------- - - -
---
JPD)P
101 JP 1 )P( 5
J1 MIii
Theoreticol relotive intensities
0.15 0.009 I 1 2.3 015 0.009
-579
-28.3 -18.5 -70 + 11.0

SI I I | I
-60 -50 -40 -30 -20 -10 0 +10 +20
3
FIGURE 7-47 The phosphorus nmr spectrum of the diphosphate anion HP20 5
and its interpretation. 1,Chemical shift differences in P(> and Pi.. II, H-P splitting.
Ill, P-P splitting. IV, Observed spectrum. [Reprinted with permission from C. F.
Callis, et al., J Amer. Chem. Soc., 79, 2722 (1957). copyright by the American
Chemical Society.]
7-23 Double Resonance and Spin-tickling Experiments 267
7-23 DOUBLE RESONANCE AND SPIN-TICKLING
EXPERIMENTS
In a double resonance experiment, the sample is subjected to a second r.f. source

whose frequency corresponds to the Larmor frequency of one of the nuclei in
the sample. This field causes the contribution to the spectrum from this nucleus
to disappear, and this nucleus is said to be decoupled. The second r.f. field is
applied with a large amplitude, at right angles to H, and orthogonal to the
pickup coils. The net effect of this field is to cause nuclear transitions in, say, B
of the AB system. Decoupling of the B nucleus is achieved, and B makes no
contribution to the A spectrum. In practice, decoupling can be accomplished
only if VA v > 5 |JABI. When the spectrum of A is examined with B being
irradiated, the symbol A -- {B} is employed to indicate this fact.
For systems where J z A, one can perform a spin-tickling experiment. In this
experiment, a weak r.f. field is employed and all transitions having an energy level
in common with the peak being irradiated will undergo a change. Referring to
Figs. 7-42 and 7-44, the I - 2 transition is seen to have an energy level in
common with the 1 -+ 3 and 2 - 4 transitions, but not the 3 - 4. Thus, the latter
peak is not split by spin tickling the 1 - 2 peak, but the other spectral lines will
be split. Experimentally, it is found that the tickling splits the lines that have an
energy level in common with the line being saturated. Furthermore, if the
transitions for the two peaks connected by a common energy level correspond
to a consecutive change in spin of both nuclei, each by 1, as in 1 - 2 and 2 - 4
(i.e., 22a to 2c#and if# to ##/3) or as in 1 -- 3 and 3 -+ 4, a sharp doublet results.
When this is not the case, as in 1 - 2 and 1 -+ 3 (i.e., 22 to a# and 2c to #lc)or
3 -+ 4 and 2 - 4, a broad doublet results. We shall not go into the reasons for
this, but simply point out that this is a valuable technique for spectral assignment
and energy level ordering in second-order systems and for determining the
relative signs of J in first-order systems.
The double resonance technique can be employed to evaluate chemical shifts
48
for nuclei other than protons by using a proton probe( ) If nucleus Y is splitting
a proton, the frequency of the r.f. field that is most effective for decoupling Y from
the protons is measured, and thus the chemical shift of Y is determined using a
proton probe. This is the basis for the INDOR technique.
The proton nmr spectrum of diborane is illustrated in Fig. 7-48(A). This
spectrum results from two sets of non-equivalent protons (bridge and terminal
protons) being split by the "B nuclei. The asterisks indicate fine structure arising
from the smaller abundance of protons on 1B nuclei. (' 01B has a natural
abundance of 18.83% and I = 3 compared to 81.17% for 'B with I = 3/2.) In
Fig. 7-48(B), the splitting caused by "B has been removed by saturation of the
boron nuclei by the double resonance technique. Two peaks of intensity ratio
4 9
2: 1 are obtained, corresponding to the four terminal and two bridge protons.( )
Two isomers have been obtained in the preparation of N 2 F2 . One definitely
has a trans structure with one fluorine on each nitrogen. In conflicting reports,
the structure of the second isomer has been reported to be the cis isomer and
also F2N=N. An excellent discussion of the results obtained by employing
several different physical methods in an attempt to resolve this problem has been
reported along with the fluorine nmr spectrum and results from a double
4
resonance experiment."') Saturation of the ' N nucleus in this second isomer
with a strong r.f. field causes collapse of all nitrogen splitting. It is concluded that
FIGURE 7-48 Proton nmr

spectrum of B2H6. (A)
Proton nmr with 11B and (
l0B splitting. (B) Proton
nmr with 11B nucleus
saturated. [From J. N.
Schoolery, Disc.
Faraday Soc., 19, 215
(1955).]
(B)
the chemical shift of the two nitrogens must be equivalent, and this eliminates
the F 2N=N structure. Additional evidence is obtained for the cis structure from
a complete spectral interpretation. The value for JN-F calculated in this study
for a cis structure is reasonable when compared to JN-F for NF .
3
A more complete discussion of the theory of the double resonance technique
and many more examples of its application are contained in a review article("'
by Baldeschwieler and Randall.
7-24 DETERMINING SIGNS OF COUPLING

CONSTANTS
The double resonance technique has been successfully employed to determine

the relative sign of coupling constants. This can be illustrated by considering the
proton nmr spectrum( 5 2 ) of(C 2 H5 )2 T-' in Fig. 7-49 (I = '/2 for 2 0 5T). If JTI-CH,
and JTI-cH, are both positive, both low field peaks correspond to interaction with
positive nuclear magnetic quantum numbers of TI. If the signs of J are different,
one low field peak corresponds to interaction with the moment from thallium
nuclei where m, = + /2 and the other to - /2. By irradiation at the center of
each of the multiplets, it was shown that each CH 3 triplet was coupled to the
JT1-Hl
"TI CH 3 )
H-c
I-CH2 (628 Hz)
Inc. field strength-

FIGURE 7-49 NMR spectrum of (C2 H,)2 TI- (facsimile).
7-25 Effects on the Spectrum of Nuclei with Quadrupole Moments 269
distant methylene quartet and vice versa. For example, irradiation with a
frequency corresponding to the low field triplet resulted in the disappearance of
the fine structure of the high field methylene signal. This result indicates that
JTI-CH, and JT 1CH2 have opposite signs. If the sign were the same, the two
low field multiplets would be coupled together as would the two high field
multiplets, and saturation of the low field triplet would cause collapse of the fine
structure in the low field methylene signal.
7-25 EFFECTS ON THE SPECTRUM OF NUCLEI

WITH QUADRUPOLE MOMENTS
Quadrupolar nuclei are often very efficiently relaxed by the fluctuating electric
fields that arise from the dipolar solvent and solute molecules. The mechanism
of quadrupole relaxation depends upon the interaction of the quadrupolar
nucleus with the electric field gradient at the nucleus. This gradient arises when
the quadrupolar nucleus is in a molecule in which it is surrounded by a
non-spherical distribution of electron density.
Thefield gradient,q, is used to describe the deviation of the electronic charge
cloud about the nucleus from spherical symmetry. If the groups about the nucleus
in question have cubic symmetry (e.g., Td or Oh point groups), the charge cloud
is spherical and the value of q is zero. If the molecule has cylindrical symmetry
(a threefold or higher symmetry axis), the deviation from spherical symmetry is
expressed by the magnitude of q. If the molecule has less than cylindrical
symmetry, two parameters are usually needed, q and q. The quantity q is referred
to as the asymmetry parameter. The word "usually" is inserted because certain
combinations of angles and charges can cause fortuitous cancellations of effects
leading to i = 0. The axis of largest q is labeled z and is described by q,.. The
other axes, described by field gradients q,, and q,,, are described by the
asymmetry parameter, which is defined as:
) = (q,, - q,,)/qzz (7-69)
The effectiveness of the relaxation depends upon the magnitude of the field
gradient. Rapid nuclear quadrupole relaxation has a pronounced effect on the
linewidth obtained in the nmr spectrum of the quadrupolar nucleus, and it also
influences the nmr spectra of protons or other nuclei attached to this quadrupolar
nucleus. In the latter case, splittings of a proton from the quadrupolar nucleus
may not be observed or the proton signal may be so extensively broadened that
the signal itself is not observed. This can be understood by analogy to the effect
of chemical exchange on the proton nmr spectra. Either rapid chemical exchange
or rapid nuclear quadrupole relaxation in effect places the proton on a nucleus
(or nuclei, for chemical exchange) whose spin state is rapidly changing. Nuclear
quadrupole relaxation rates often correspond to an intermediate rate of chemical
exchange, so extensive broadening is usually observed. As a result of quadrupole
relaxation, the proton nmr spectrum of 14NH 3 ('4 N, I = 1) consists of three very
broad signals; while in the absence of this effect, the spectrum of "NH 3 ( 5 N,
I = 1/2) consists of a sharp doublet. On the other hand, in '4 NH 4 , where a
spherical distribution of electron density gives rise to a zero field gradient, a
sharp three-line spectrum results. In a molecule with a very large field gradient,
a broad signal with no fine structure is commonly obtained.
When one attempts to obtain an nmr spectrum of a nucleus with a
quadrupole moment (e.g., 3 5C and 4 N) that undergoes relaxation readily, the
signals are sometimes broadened so extensively that no spectrum is obtained.
This is the case for most halogen (except fluorine) compounds. Sharp signals
have been obtained for the halide ions and symmetrical compounds of the
halogens (e.g., ClO, ), where the spherical charge distribution gives rise to only
small field gradients at the nucleus, leading to larger values for T,.
Solutions of I- (127 I = /2) give rise to an nmr signal. When iodine is
added, the triiodide ion, 13-, is formed, destroying the cubic symmetry of the
iodide ion so that quadrupole broadening becomes effective and the signal
disappears. Small amounts of iodine result in a broadening of the iodide
resonance, and the rate constant for the reaction I- + 12 - 13 can be calculated
broadening.!s 3 > It is interesting to note that chlorine chemical shifts have been
observed 54 ) for the compounds: SiCl4 , CrO 2 Cl 2 , VOCI3 , and TiCI4 , where the
chlorine is in an environment of lower than cubic symmetry.
An interesting effect has been reported for the fluorine nmr spectrum of
NF 3 . The changes in a series of spectra obtained as a function of temperature
are opposite to those normally obtained for exchange processes. At -205 C a
sharp single peak is obtained for NF 3 ; as the temperature is raised the line
broadens and a spectrum consisting of a sharp triplet (I = 1 for 'N) results at
20 C. It is proposed that at low temperature the slow molecular motions are
(A) most effective for quadrupole relaxation of 4 N; as a result, a single line is
obtained. At higher temperatures, relaxation is not as effective and the lifetime
of a given state for the 4 N nucleus is sufficient to cause spin-spin splitting. A
similar effect is observed for pyrrole. 5 )' The 'N spectrum of azoxybenzene
exhibits only a singlet. The nitrogens are not equivalent, and it is
0
QN=NQ
proposed that the field gradient at the N-O nitrogen is so large as to make
this resonance unobservable.
The double resonance technique has been successfully used on the proton
nmr spectrum of AI(BH 4)3 . This molecule contains six Al-H-B bridge bonds.
(C)
Both B and 27 A1(I = '/2) have quadrupole moments. The proton nmr at 30 MHz
FIGURE 7-50 Proton nmr of
consists(5 6) of a single broad line (Fig. 7-50(A)). When the " B nucleus is saturated
AI(BH4)3 . (A) Proton
(1H -- {"B}), the proton resonance spectrum in Fig. 7-50(B) results. Fig. 7-50(C)
resonance. (B) Proton
resonance, 11B saturated.
represents the proton nmr spectrum when the sample is irradiated with frequency
(C) Proton resonance, 2 7Al corresponding to that of 2 7Al ('H - {2 7 Al}). The four large peaks in (C) arise
saturated. [From R. A. Ogg, from "B splitting of the proton and the smaller peaks from 0 B splitting. The
Jr., and J. D. Ray, Disc. bridging and terminal hydrogens are not distinguished because of a rapid proton
Faraday Soc., 19, 239 (1955).] exchange reaction that makes all hydrogens magnetically equivalent.
1. H. S. Gutowsky, "Physical Methods of Organic Chemistry," 3rd Ed., part 4, A. REFERENCES

Weissberger, ed. (Vol. 1 of "Techniques of Organic Chemistry"), Interscience, New
York, 1960.
CITED
2. J. A. Pople, W. G. Schneider, and A. J. Bernstein, "High Resolution Nuclear Magnetic
Resonance," McGraw-Hill, New York, 1959.
3. (a) J. W. Emsley, J. Feeney, and L. H. Sutcliffe, "High Resolution Nuclear Resonance
Spectroscopy," Vol. 1, Pergamon Press, New York, 1966.
(b) Same authors and title; Vol. 2.
4. M. W. Dietrich and R. E. Keller, Anal. Chem., 36, 258 (1964) and references therein.
5. N. S. Bhacca. L. F. Johnson, and J. N. Shoolery, "NMR Spectra Catalog," Varian
Associates, Palo Alto, Calif., 1962; "The Sadtler Standard Spectra; NMR," Sadtler
Research Laboratories, Philadelphia, 1972; L. F. Johnson and W. C. Jankowski, "C-13
NMR Spectra" (a collection), Interscience, New York, 1972.
6. L. Eberson and S. Farsen, J. Phys. Chem., 64, 767 (1960).
7. A. L. Van Gelb, Anal. Chem., 42, 679 (1970).
8. (a) J. W. Emsley et al., "High Resolution Nuclear Magnetic Resonance Spectroscopy,"
Vol. 1, Pergamon Press, New York, 1966;
(b) P. Diehl, E. Fluck, and R. Kosfeld, eds., "NMR Basic Principles and Progress,"
Vol. 5, Springer-Verlag, New York, 1971.
9. E. Fluck, "Anorganische und Allgemeine Chemie in Einzeldarstellungen. Band V. Die
Kernmagnetische Resonanz und Ihre Anwendung in der Anorganische Chemie,"
Springer-Verlag, Berlin, 1963.
10. A. D. Buckingham and P. J. Stephens, J. Chem. Soc., 1964, 4583.
11. C. E. Johnson and F. A. Bovey, J. Chem. Phys., 29, 1012 (1958) and references therein.
12. (a) A. D. Buckingham, Can. J. Chem., 38, 300 (1960);
(b) J. I. Musher, J. Chem. Phys., 37, 34 (1962).
13. W. N. Lipscomb et al., J. Amer. Chem. Soc., 88, 5340 (1966); Adv. Mag. Res., 2, 137
(1966).
14. D. Gurka and R. W. Taft, J. Amer. Chem. Soc., 91, 4794 (1969).
15. F. Prosser and L. Goodman, J. Chem. Phys., 38, 374 (1963).
16. (a) M. Karplus and T. P. Das, J. Chem. Phys., 34, 1683 (1961);
(b) R. J. Pugmire and D. M. Grant, J. Amer. Chem. Soc., 90, 697 (1968).
17. W. H. Pirkle, S. D. Beare, and R. L. Muntz, J. Amer. Chem. Soc., 91, 4575 (1969);
ibid., 93, 2817 (1971) and references therein.
18. H. S. Gutowsky, J. Chem. Phys., 37, 2136 (1962).
19. J. S. Waugh and F. A. Cotton, J. Phys. Chem., 65, 562 (1961).
20. R. A. Hoffman and S. Gronowitz, Arkiv. Kemi., 16, 471, 563, (1961); 17, 1 (1961);
Acta Chem. Scand., 13, 1477 (1959).
21. M. Karplus, J. Chem. Phys., 33, 1842 (1960).
22. (a) M. T. Rogers and J. D. Graham, J. Amer. Chem. Soc., 84, 3666 (1962);
(b) J. D. Kennedy et al., Inorg. Chem., 12, 2742 (1973).
23. H. S. Gutowsky, D. W. McCall, and C. P. Slichter, J. Chem. Phys., 21, 279 (1953).
24. C. F. Callis, J. R. Van Wazer, J. N. Shoolery, and W. A. Anderson, J. Amer. Chem.
Soc., 79, 2719 (1957).
25. R. 0. Ragsdale, and B. B. Stewart, Inorg. Chem., 2, 1002 (1963).
26. N. Muller, J. Chem. Phys., 36, 359 (1962).
27. H. S. Gutowsky, and C. Juan, J. Amer. Chem. Soc., 84, 307 (1962); J. Chem. Phys.,
37, 2198 (1962).
28. J. H. Goldstein and G. S. Reddy, J. Chem. Phys., 36, 2644 (1962).
29. R. S. Drago and N. A. Matwiyoff, J. Chem. Phys., 38, 2583 (1963); J. Organomet.
Chem., 3, 62 (1965).
30. H. A. Bent, Chem. Rev., 61, 275 (1961).
31. J. Hinze and H. H. Jaffe, J. Amer. Chem. Soc., 84, 540 (1962).
32. M. Karplus, J. Amer. Chem. Soc., 85, 2870 (1963).

33. S. Sternhall, Quart. Rev. (London), 23, 236 (1969) and references therein.
34. M. A. Cooper and S. L. Manatt, J. Amer. Chem. Soc., 92, 1605 (1970).
35. R. L. Middaugh and R. S. Drago, J. Amer. Chem. Soc., 85, 2575 (1963).
36. J. R. Holmes and H. D. Kaesz, J. Amer. Chem. Soc., 83, 3903 (1961).
37. N. A. Matwiyoff and R. S. Drago, Inorg. Chem., 3, 337 (1964).
38. G. D. Shier and R. S. Drago, J. Organomet. Chem., 6, 359 (1966).
39. T. F. Bolles and R. S. Drago, J. Amer. Chem. Soc., 88, 5730 (1966).
40. J. F. Nixon and A. Pidcock, Ann. Rev. NMR Spectroscopy, 2, 345 (1969).
41. R. K. Harris, Canad. J. Chem., 42, 2275, 2575 (1964).
42. (a) B. E. Mann, B. L. Shaw, and K. E. Stainbank, Chem. Comm., 151 (1972);
(b) D. E. Axelson and C. E. Holloway, Chem. Comm., 455 (1973).
43. E. L. Muetterties, J. Amer. Chem. Soc., 82, 1082 (1960).
44. E. L. Muetterties, W. Mahler, and R. S. Schmutzler, Inorg. Chem., 2, 613 (1963).
45. J. D. Roberts, "An Introduction to the Spin-Spin Splitting in High Resolution NMR
Spectra," W. A. Benjamin, New York, 1961.
46. P. L. Corio, Chem. Rev., 60, 363 (1960).
47. C. F. Callis et al., J. Amer. Chem. Soc., 79, 2719 (1957).
48. G. C. Levy and G. L. Nelson, "Carbon-13 Nuclear Magnetic Resonance for Organic
Chemists," Wiley-Interscience, New York, 1972.
49. J. N. Shoolery, Disc. Faraday Soc., 19, 215 (1955).
50. J. H. Noggle, J. D. Baldeschwieler, and C. B. Colburn, J. Chem. Phys., 37, 182 (1962).
51. J. Baldeschwieler and E. W. Randall, Chem. Rev. 63, 82 (1963).
52. J. P. Maher and D. F. Evans, Proc. Chem. Soc., 1961, 208; D. W. Turner, J. Chem.
Soc., 1962, 847.
53. 0. E. Myers, J. Chem. Phys., 28, 1027 (1958).
54. Y. Masuda, J. Phys. Soc. Japan, 11, 670 (1956).
55. J. D. Roberts, J. Amer. Chem. Soc., 78, 4495 (1956).
56. R. A. Ogg, Jr., and J. D. Ray, Disc. Faraday Soc., 19, 239 (1955).
COMPILATIONS Proton Shifts

N. S. Bhacca, L. F. Johnson, and J. N. Shoolery, "NMR Spectra Catalog," Varian
OF CHEMICAL Associates, Palo Alto, Calif., 1962.
"The Sadtler Standard Spectra; NMR," Sadtler Research Laboratories, Philadelphia,
SHIFTS 1972.
N. F. Chamberlain, "The Practice of NMR Spectroscopy with Spectra-Structure Correla-
tions for Hydrogen-i," Plenum, London and New York, 1974.
3
1C Shifts
L. F. Johnson and W. C. Jankowski, "C-13 NMR Spectra," Wiley-Interscience, New York,
1972.
G. C. Levy, R. L. Lichter, and G. L. Nelson, "Carbon-13 Nuclear Magnetic Resonance
Spectroscopy," 2nd ed., Wiley-Interscience, 1989.
G. C. Levy, "Topics in 3 C NMR Spectroscopy," Wiley-Interscience, New York, volumes
from 1974 to present.
B. E. Mann and B. F. Taylor, " 3 C NMR Data for Organometallic Compounds,"
Academic Press, New York, 1981.
Q. T. Pham, R. Petiaud, and H. Waton, "Proton and Carbon NMR of Polymers," Wiley,
New York, 1983.
Exercises 273
P Shifts
E. Fluck, "Anorganische und Allgemeine Chemie in Einzeldrstellungen. Band V. Die
Kernmagnetische Resonanz und Ihre Anwendung in der Anorganische Chemie,"
Springer-Verlag, Berlin, 1963.
"Methods in Stereochemical Analysis," Volume 8, by J. G. Verkade and L. D. Quin, eds.,
VCH Publishers, Deerfield Beach, Fla.
" 3 1 P NMR, Principles and Applications," D. G. Gorestein, ed., Academic Press, New
York, 1984.
"F Shifts
C. H. Dungan and J. R. VanWazer, "Compilation of Reported "F Chemical Shifts,"
Wiley-Interscience, New York, 1970.
A. J. Gordon and R. A. Ford, "The Chemists Companion," Wiley, New York, 1972.
"B Shifts
H. Noth, "NMR Spectroscopy of Boron Compounds," Springer-Verlag, New York, 1978.
Miscellaneous Nuclei Shifts

"NMR of Newly Accessible Nuclei." P. Laszlo, ed., Academic Press, New York, 1983.
R. K. Harris and B. E. Mann, "NMR and the Periodic Table," Academic Press, New
York, 1978.
General Text
F. A. Bovey (with L. Jelinski and P. A. Mirav) "Nuclear Magnetic Resonance Spec-
troscopy," 2nd ed., Academic Press, New York, 1988.
For more practice with Spectral Interpretation

R. M. Silverstein, G. C. Bassler, and T. C. Morrill, "Spectrometric Identification of Organic
Compounds," 4th ed., Wiley, New York, 1981.
1. The b value of a substance relative to the external standard. methylene chloride, is EXERCISES
-2.5. Calculate J relative to external standards (a) benzene and (b) water using the
data in Appendix F.
2. Assuming the relationship discussed between Jis-- and hybridization, what would
be the ratio of the coupling constants in a five-coordinate and six-coordinate complex
of (CH 3 )3 SnCl [i.e., (CH3) 3 SnCI - B and (CH3) 3 SnCl - 2B]?
3. The compound B[N(CH 3 )2]3 is prepared and dissolved in a wide number of different
solvents. Propose a method of determining in which ones the solvent is coordinated
to the compound.
4. Consider the diamagnetic complex (Me 3 P)4 Pt". Sketch the phosphorus resonance
signal if
a. JP-H > JP-Pt-
b. JPPt > JPH.
5. Indicate the number of isomers for cyclic compounds of formulas P 3N 3(CH 3)2 C 4,
and sketch the phosphorus resonance spectrum of each (assume A > J, JP--H is small,
and JP-H can be ignored for phosphorus atoms which do not contain methyl groups).
6. Would you expect the 14 N nmr spectrum to be sharper in NH 3 or NH 4 '? (For "'N,
I = 1.) Explain.
7. Consider all possible isomers that could be obtained for the eight-membered ring
compound P 4 N 4 Cl6 (NHR)2 and indicate the ideal phosphorus resonance spectrum
expected for each. Which of the above are definitely eliminated if the phosphorus
resonance consists of two triplets of equal intensity?
8. The proton nmr spectrum of S-O is not a singlet. Is the SO3 group
CH2-O
planar? What would the spectrum look like if the sulfur underwent rapid inversion?
9. It is found that the methylene groups in (CH 3CH 2 )2 SBF 3 give rise to a single
methylene resonance. Explain.
H 0
10. a. Are the methyl groups in (CH 3) 2C-P-C,H equivalent?
Cl
b. Ignore the splitting of the methyl groups by the phenyl protons in the above
compound and assume J < A. What would the spectrum of the methyl protons
look like?
11. The proton nmr spectra of a series of compounds are given below. Assign their
geometries and interpret the spectra.
100 MHz
H- 10-0 9.0 8.0 ppm
60 MHz
(2)] TMS
(1)
12.0 10.0 8.0 6.0 4.0 2.0 0.0 -2.0 -4.0 ppm
Exercises 275
a. The compound is ReCI 3[P(CH 3)2 CH 5 ] 3 . Curve (1) is the full spectrum, (2)
is a more intense sweep of the 6.0 to 13.0 ppm region, and (3) is the 100 MHz
spectrum of the 7.0 to 10 ppm region.
b. Why is the spectrum in part (a) at 100 MHz different from that at 60 MHz?
Why are the chemical shifts in ppm relative to TMS the same?
c. A compound with empirical formula C4 H1 1 N.
2.6 0.8 ppm
d. A compound with empirical formula C 2 H3 F 3 0.
ppm
e. A compound with empirical formula C2 H 4 0.
9.8 1.6 PPM

f. A compound with empirical formula C4 H,.
5.9 2.2 ppm
g. A compound with empirical formula C H 02-

4 8
ppm
Relative
areas
h. A compound with empirical formula C 3 HCIF2 -
J =18
J=12
1.7 0 ppm
3.6 1.7 0 ppm
Exercises 277
i. A compound with empirical formula C 8 H,,O. The peak at 2.4 ppm vanishes
in D 2 0.
(nondescript multiplet)
7.2 3.7 2.8 2.4 ppm
5 2 2 1 Relative
area
j. A compound with empirical formula CHiO 4 P.
8. 7.0 6.0 5.0 4.0 30 201 1.0 0

8.0 7.0 6.0 1 5.o0 1 4.0 j 3.0 1 2.0 1.0 0.0 ppm
k. A compound with empirical formula C 3 H5 0 2C.
12.0 4.9 1.5 ppm

1. A compound with empirical formula C 2 HNCI dissolved in water.
8.4 3 3.7 1.7 ppm
m. A compound with empirical formula C 1 4 H 22 0 4 .
I I i
4.5 2.0 1.3
n. A compound with empirical formula (SiH 3 )2 PSiH 2 CH 3 .
0 Hz 321 Hz
Exercises 279
o. A compound with empirical formula Pt[P(C 2 H5 )3] 2 HCl. (The peaks at 8 ppm
are nondescript multiplets.)
8 15.6 26.9 38.2 ppm

72 108 1 4 1 Relative
areas
p. A compound with empirical formula C,,H,,. The curves above the peaks
represent the integrated intensities of the peaks. The relative areas can be obtained
by comparing the heights (number of squares) of the respective integration curves.
7.0 5.0 1.0 0 ppm
q. A compound with empirical formula CH,,O. The curves above the peaks
by comparing the heights, in squares, of the respective integration curves.
ppm
r. A compound with empirical formula CHioO. The curves above the peaks
by comparing the heights, in squares, of the respective integration curves.
Hi!!!!!HHiHH!! !.!!!!i ! ii i!i i ii !
3.0 2.0 1.0 0.0

ppm
s. A compound with empirical formula CsHOF.

Exercises 281
t. The "C nmr of OSC 2 D, at 75.1 MHz (D = deuterium with a spin of 1).
39.01-
39.64
F- 38.73
40.117
-38.45
55 50 45 40 35 30 25
ppm
12. Spectrum (a) results for the methyl region of one isomer of Pd[P(CH 3)2C6 H5 ] 2 12 -
Spectrum (b) is obtained in the methyl region for the other isomer of
Pt[P(CH 3)2 C6 H5 ] 2 12 .
1 2 1 Relative areas
4 1
Related areas
a. Which is cis and which is trans?
b. Explain why the areas in (b) are in the ratio of I to 4 to I.

13. The 'B nmr spectrum below is obtained for 5-C-2,4-C 2 B5 H,. Interpret the
spectrum. (Hint: B7 H7 2 - is a pentagonal bipyramid of B-H subunits in which the
axial borons are numbered 1 and 7.)
14. The 9 F nmr spectrum of PF 3 (NH 2 )2 is reported in (a), and expanded spectra of the
individual peaks are presented in (b) through (e). Propose a structure and interpret
the spectrum.
(A)
42.9 ppm 59.7 ppm 66.6 ppm 88.1 ppm
20 cps
(B)
42.9 ppm
(C)
59.7 ppm
40 cpsa
(D)
66.6 ppm
(E )
88.1 ppm
Exercises 283
15. The proton resonance of the hydridic hydrogen of HNi[OP(C2H,) 3]4 is given as
follows. Propose a structure for this complex.
26 Hz
16. Given the following proton nmr spectra and molecular structures, assign all peaks:
(a)
(a) Me (C)
Me-, CH so(--)
Me CH2 CH2 Na
(a) (b) (d)
I II I I I I I II I I I I
8.0 7.0 6.0 5.0 4.0 3.0 2.0 1.0 0.0
ppm
(a)
Me
H
(b)
I | I lI I i I II I I I
8.0 7.0 6.0 5.0 4.0 3.0 2.0 1.0 0.0

ppm
(b)
H
H (c)
(b)H C) H(d)
(c)H Fe / - H(e)
(d)H --
@1 H(eNO
2
I I I I I I I I I I I I I I - I-
8.0 7.0 6.0 5.0 4.0 3.0 2.0 1.0 0.0

ppm
17. What properties in a functional group X (which can consist of more than one atom)
attached to an isopropyl group can cause non-equivalent methyls to occur in the
isopropyl group?
18. The following proton nmr spectrum is that of a compound with the molecular formula
CH, NO. Propose a structure consistent with the formula and spectrum. Insofar as
possible, assign the peaks. Numbers in parentheses refer to relative peak areas.
TMS
I I I I I I I I I I I I I I
8.0 7.0 6.0 5.0 4.0 3.0 2.0 1.0 0.0

(4) (2) (2) (3) ppm
19. Interpret the spectrum below for the compound CH 3 Pt(CH 2 =CH2 )[PC6 H ,(CH 3)2] 2 .
31
Justify all splittings. Is this the cis or trans isomer? (Hint: For p, I = 1/2, abun-
9 34
dance = 100%. For ' Pt, I = '/2, abundance = %.) [See Inorg. Chem., 12, 994
(1973).]
Exercises 285
5 = -1.81 ppm
6= - 4
.12 ppm
6 = -0.83 ppm
rei. 4 16 4 1248 12 3 12 3
areas
20. Sketch the spectra for the following molecules and conditions (assume 100% abun-
dances):
H
|7 F
a. H 2 PF 3 , structure F-Ps JP-F >JP-H IJH-F 'H 2
i F IF 2
H 12
9
H, 1 F, and 'P resonances
b. N(CH 2 F)4 * JFCH FCN' JFCNCH 0

19
F only
21. The following proton nmr spectrum is reported [Inorg. Chem., 2, 939 (1963)] for
((NH2) 3P) 2 Fe(CO) 3 (the two phosphorus ligands are trans).
Tdp 2 Fe(CO) 3
a. Explain the splittings, assuming large P-P coupling, and that there is no proton-
nitrogen coupling observed.
b. Why isn't the N-H coupling observed?
c. Using the A, B, X designation, label this system.

22. The spectrum of the following compound is a complex type. Explain how the double
resonance technique could be employed to aid in interpreting the spectrum.
H H
"N /
C--C
// \\
H-C, C-H
23. The spectrum in the following figure is that of an AB-type of molecule.
120cps
18OcpsI
a. Calculate J and A.
b. A 40-MHz probe was employed. Calculate the difference in 6 for the two peaks.
c. If e occurs at a 6 value of 2.8, what are the 6 values for A and B?
24. Consider the series i-propyl X (where X = Cl, Br, I).
a. In which compound would the remote paramagnetic effect from the halogen be
largest? Why?
b. In which would the remote diamagnetic effect be largest?
25. What would the nmr spectrum of PF, look like under the following conditions
(AF(a)-F(b) > JF(a)-F(b)
a. Very slow fluorine exchange.
b. Rapid intermolecular fluorine exchange.
c. Rapid intramolecular fluorine exchange.
26. Consider the molecule shown in problem 8. Using the A, B, X, ... terminology, classify
this molecule and indicate the non-equivalent protons.
27. a. List and briefly describe in your own words the factors that influence the
magnitudes of proton chemical shifts.
b. Why do 9 F and "C chemical shifts cover a much larger range than do those of
protons?
28. Ramsey's formula is used to calculate the local contributions to the chemical shielding
of a nucleus; it is given as equation (7-37). The nuclear magnetic resonance spectrum
of 59Co has been observed in a variety of environments. A correlation has been
proposed relating the chemical shift to the wavelength of an electronic absorption
observed in these complexes.
Exercises 287
Complex ; (Angstroms)
K 3 [Co(CN)6] 3110
[Co(en) 3 ]Cl3 4700
[Co(NH 3)6]C 3 4750
(en = ethylenediamine)
a. Do you expect the diamagnetic term in Ramsey's formula to account for an 8150
ppm variation in the chemical shift observed in this series of complexes? Explain
briefly.
b. Do you expect the paramagnetic contribution to the chemical shielding to vary in

this series of complexes? Why?
c. List these cobalt complexes in order of increasing magnetic field for resonance.
Give your reasoning.
29. Ramsey's equation [equation (7-37),] is used to calculate local contributions to

chemical shifts. Consider the following boron compounds and their "B chemical
shifts (in ppm relative to BF 3 etherate).
Three-coordinate Four-coordinate
B13(-11 = +5.5 NaBF= +2.3

BF = -94 B(OH) 1.8
BCI3 = - 47.7 B, piperidine +2.3
BBr 3 = 40.1 NaB(C 6H5 ) 4 =8.2
B(OCH 3 )3 = - 18.1 LiB(OCH 3 )4 2.9

B(C2 H5 )3 = -85.0 BF3 - P(CH 5 )3 = -0.4
The generalized m.o. description for three-coordinate boron compounds consists of

bonding, non-bonding (empty boron a.o.), and antibonding molecular orbitals (I).
Four-coordinate boron compounds are described by bonding and antibonding m.o.'s
(no non-bonding m.o.) (II).
U* I*
a. Can the first term of the Ramsey equation explain the "B chemical shifts of the
three-coordinate compounds? Why or why not?
b. Rationalize from the m.o. description the fact that the "B chemical shift of the
three-coordinate complexes varies 125 ppm, while that of the four-coordinate
species varies 11.1 ppm.
30. It was said that there is no contribution to the local paramagnetic shielding in a
linear molecule aligned parallel to the applied magnetic field. Construct molecular
orbitals for the molecule HF and show that matrix elements resulting from the Ramsey
equation lead to the foregoing result.
31. Consider benzene and cyclohexane.
a. In which do you think the proton resonances would be further upfield? Why?
b. In which do you think the "C resonances would be further upfield? Why?
3
c. In which would J1C-H be larger? Why?
199
32. In compounds of the type CH 3HgX, the Hg- 2H coupling constant is observed to
vary by more than a factor of two, depending on the substituent X. Some examples
are given below:
X J(Hz)
CH 3 104
I 200
Br 212
Cl 215
CIO 4 233
Propose an explanation for the observed splittings.
33. A number of ' 5 N-labeled aminophosphines have been synthesized and the 15N-H
nmr coupling constants have been measured. A sample of the results is as follows:
Compound J(Hz)
F4 P 15 NH 2 90.3
(CF 3)2 P 15 NH 2 85.6
F 2 P"NH2 82.7
There are two possible explanations in terms of the bonding in this series.
a. Give the explanation that is based on the extent of N - P n-bonding.
b. Offer another explanation that was presented, which also accounts for these results.
34. a. Would you expect the difference 6CH3 -- CH in C2 H51 to be independent of the
remote anisotropy in the C-i bond? Why?
b. Will this difference have a greater or smaller contribution from anisotropy than
6CH2?
c. If in a series of compounds the CH2 values showed a different trend than the
6CH3 - 6CH, values, what could you conclude about anisotropic contributions?
Exercises 289
35. Combine in your thinking (reread if necessary) the discussion on interatomic ring
currents with the relationship between J C-H and T (discussed in section 7-18 on
Applications of Spin-Spin Coupling to Structure Determination). Using these con-
cepts, propose a method for determining if there is anisotropy in the methyl
chemical shift of B-trimethyl borazine.
8 Dynamic and
Fourier Transform NMR
Introduction In this chapter we consider the applications of nmr to dynamic systems and then
proceed to a discussion of Fourier transform nmr. The chapter concludes with
a discussion of liquid crystal and solid nmr. The nmr spectra of paramagnetic
ions (Chapter 12) is discussed after some fundamental background in magnetism
is developed (Chapter 11).
Evaluation of As mentioned in Chapter 7, when two species undergo rapid exchange on the
nmr time scale, the chemical shift observed is a mole-fraction weighted average
Thermodynamic of the two resonances. With rapid exchange:
Data with NMR
A+ X : 2AX (8-1)
the chemical shift of the A (or X) resonance will be a mole-fraction weighted

average of the resonance of A (or X) and that of the analogous atom in the AX
adduct, as shown in equation (8-2).
6obs(A) = NA6A + NAX6AX (8-2)

N refers to mole fraction. An analogous equation could be written for the
resonance, obs,(X).
This exchange averaging can be used(') to evaluate the equilibrium constant
for a reaction. The approach will be illustrated by deriving an expression for the
equilibrium constant of the reaction illustrated in equation (8-1). Assume that
we are observing the chemical shift of a proton of molecule A that is shifted
considerably upon forming the complex AX and that there is rapid exchange
between A and AX giving a mole-fraction weighted shift as in equation (8-2).
Expressing equation (8-2) in molarity units, we obtain for the reaction in
equation (8-1):
290 obs - [A] [AX]

[A] + [AX] A [A] + [AX] AX
Rearranging, collecting terms, and subtracting [AX]6A from both sides of
the equation produces:
[A](6Obs -- A) + [AX](6.b, - 6A) [AX](6AX - 6A)
Defining A6., as (o.bs -- A) and A6CA as (JAI -A), we can write:
[AX] = E A6Obs (8-3)

A6CA
where [A ], the initial concentration of A, equals [A] + [AX].

Substituting equation (8-3) into the equilibrium constant expression
K- [AX]
([A ] - [AX])([X ] - [AX])
one obtains equation (8-4)
K = A"bs (8-4)
(A 6CA- A6 o) XI A AOb [A ](8 4
cA
In equation (8-4), all quantities are known except K and A3 CA. The two unknowns
are constant at a given temperature and can be obtained" by solving a series
of simultaneous equations that result from measuring A6oss in a series of
experiments in which [X ] and [A ] are varied. This aspect of the problem is
similar to that described in Chapter 4 for systems that obey Beer's law.
8-1 RATE CONSTANTS AND ACTIVATION ENTHALPIES NMR Kinetics

FROM NMR
The proton nmr spectrum of a deuterated cyclohexane molecule, (shown in Fig.

8-1(A)), as a function of temperature is shown in Fig. 8-1(B) to (G). Two isomers
are possible where the hydrogen is in either the axial or equatorial position of
the cyclohexane ring. At room temperature, the two forms are rapidly intercon-
verting and there is no contribution to the observed line width from exchange
effects. As the temperature is lowered, the band begins to broaden from chemical
exchange contributions to the nuclear excited state lifetime. This range, down to
the temperature at which the two separate peaks are just beginning to be resolved,
is referred to as the near fast exchange region. As the temperature is lowered
further, two peaks are seen and their chemical shifts change as a function of
temperature until curve (F) is obtained. This range is referred to as the intermediate
exchange region. As the temperature is lowered further, the chemical shifts of the
peaks no longer change, but the resonances sharpen throughout the so-called
slow exchange region. Finally, at -79 and lower, there are no kinetic contribu-
292 Chapter 8 Dynamic and Fourier Transform NMR
Fast
exchange
250
Near
fast
exchange
Coalescence
Intermediate
exchange
(F)
Slow
exchange
(G)
-790 Stopped
exchange
FIGURE 8-1 Deuterated cyclohexane, C6D,,H, (A) and the

temperature dependence of its nmr spectrum (B to G).
tions to the shape of the spectrum, and this is referred to as the stopped exchange
region.
By a full analysis of the influence of chemical exchange on the magnetization
via the Bloch equations, it is possible to derive equations for the evaluation of
rate constants from the nmr spectra. The derivations are beyond the scope of
this treatment and the reader is referred to Emsley, Feeney, and Sutcliffe,
Volume 1, Chapter 9, or Pople, Schneider, and Bernstein, p. 218. for more details.
We shall present the results of these derivations in a form useful for a kinetic
analysis and comment on shortcomings of the various approaches that have been
employed.
One of the simplest systems is one in which a given proton can be at either
one of two molecular sites; the probability that it will be at one site is equal to
the probability that it will be at the other, and it has the same lifetime at each.
The cyclohexane interconversion and many other isomer or rotamer exchange
problems satisfy these criteria. If one works with the chemical shift changes, crude
rate data can be extracted from any two-site (A and B) equal-lifetime process in
the intermediate exchange region by using equations (8-5).
(vA - VB)s, _ 1 1/2

2
VA
0 -VB
0
L
2
7'T
A0
-V
B2 (8-5)
- vBO) is the separation of peaks in the stopped exchange

0
In this equation (VA
region in Hz; (VA - vB)oS is the separation of peaks at a particular temperature
in the intermediate exchange region. Recall that we stipulated that the two
lifetimes, TA' and TB', are equal. The lifetime T' then is the sum of the two, or the
lifetime at site A is simply T'/2.
At the coalescence temperature, the two peaks have just merged, so that
(VA - vB)obS equals zero. The approximations used to derive equation (8-5) no
longer apply, and equation (8-6) is used to obtain the lifetime:
0
- v ) (8-6)
.' 2n(v
Equation (8-6) shows that the necessary condition for detecting two
exchanging nuclei as separate resonances is given by:
r' > 2 /2 (8-7)

2nr(vA' - vB0)
Thus, the farther apart the chemical shifts at sites A and B, the shorter the lifetime
or the faster the kinetic process will have to be to average them. Recall the
discussion of this effect in Chapter 4.
Utilization of the foregoing equations [especially equation (8-5)] leads to
rate constants with large error limits. Activation enthalpies are obtained from
the temperature dependence of the rate constant via the Arrhenius equation.
Since the temperature range corresponding to the intermediate exchange region
usually is very narrow, huge errors in the activation enthalpies can result.
Considerably more accurate information is available about the kinetic process
by using line width changes over the entire temperature region that the spectrum
is influenced by the kinetic process. These regions are illustrated in Fig. 8-2,
where a typical plot is shown for the change in the full width of the resonance
line at half height as a function of I/T
In the slow exchange region, the observed line width, 1/T 2A', of a given peak
for site A has contributions to it from the natural line width, I/T2A, and from
chemical exchange, TA, according to:
1 1 1
nAvl/ 2(A) + - (8-8a)
T2 A' T2 A TA
Since this equation employs units for Av of Hz, Av1 /2 represents the full line width
at half height.
This equation and all those in the equation (8-8) series are for a two-site
problem with equal lifetimes at both sites, but not necessarily equal populations.
This equation does not apply if there is spin-spin coupling of the protons, for
contributions of this process to 1/T2 are not included. The 1/T2 A contribution to
the width in the slow and intermediate exchange regions can be determined by
extrapolating the line for no exchange effects (dotted line a). By difference,
(1/T2 A 1 T2 A), one obtains 1/TA. The same calculation could be carried out
with the B resonance.
In the near fast exchange region, the contributions to the line width are
given by
1 NA NB
- + + NA2 NB 2 A BO (A + TB) (8-8b)
T2A' 2A 2B
where WA = 2 7TVA, and WA are the chemical shifts in the absence of

BOand
exchange. The first two terms on the right-hand side of the equality are
Peak
coalescence
FIGURE 8-2 The influence

of chemical exchange on the Near fast Slow and intermediate
line width, Av,,,, of a spectral log AV1/2 exchange exchange
band as a function of
absolute temperature, T.
(The log is plotted to give Fast
linear activation energy plots exchange
for various regions.) -- N exchange
8-2 Determination of Reaction Orders by NMR 295
mole-fraction weighted averages of 1/T2A and l/T2B for the A and B sites in the
absence of exchange. The last term is the exchange-broadened contribution. The
quantities NA/2A and NB/T 2 Bcan also be obtained by virtue of the fact that in
the fast exchange region
1 _NA N
nr Av 1 /2 - I=A A + B (8_8c)
T2 A' T2A T2.
Because one can work over a larger temperature range with the line width
equations, one can obtain more accurate values for the activation enthalpies than
with equations 8-5 and 8-6.
Neither of the above approaches (chemical shifts or line widths) are used (B)
any more to evaluate rate constants. Instead, the entire spectral line shape would
be calculated using reported equations.12 ) One can then obtain the exchange rate
by varying r until the calculated line shape fits the experimental curve. This
analysis is generally carried out by a computer that calculates the line shape (C)
using an estimated value of c, i.e., one from a rough Avl 2 calculation. The
difference between the calculated and experimental intensities is determined. The
computer then varies T until the difference is minimized. Thus, from the line shape
at a given temperature, the rate constant can be determined and the activation
(D)
parameters obtained in the usual manner from the temperature dependence of
the rate constant. A typical comparison between a calculated and an experimental
spectrum is illustrated in Fig. 8-3. The spectrum 2 ) is that of the methyl proton
of 2-picoline. The picoline is undergoing exchange with Co(2-pic)2 Cl 2 in
(E)
(CD 3)2 CO solvent and the region is the near fast exchange.
When one wishes to study a multisite problem or when there is spin-spin
coupling to one of the protons involved in the exchange, a better method involves
use of the density matrix approach.(3 5) This approach is beyond the scope of our FIGURE 8-3 Comparison of
treatment and the reader is referred to references 3 to 5 for details. experimental (solid line) and
theoretical (dots) nmr methyl
resonance spectra at
T= -44 . All solutions are
8-2 DETERMINATION OF REACTION ORDERS BY NMR 0.09 M in Co(2-pic) 2Cl2 and
contain the following
We return now to a more complete discussion of the T' values in order to concentrations of excess
understand how to obtain information about reaction orders. In the nmr 2-picoline: (A) 3.54 M, (B)
2.64 M, (C) 2.40 M, (D) 2.11 M,
experiment, we are studying a reaction occurring at chemical equilibrium. A
(E) 1.85 M. [From S. S.
decay in the net magnetization of the sample results as the forward and reverse Zumdahl and R. S. Drago, J.
reactions occur. Suppose, for example, that the decay at site A is observed. The Amer. Chem. Soc., 89, 4319
magnetization decays by a first-order decay process just like the decay of (1967).]
radioactive material. The rate constant, 1/-A, is a first-order rate constant for the
decay of the initial number of protons in our sample at site A, i.e.,
d[A] 1
- = - [A]
dt TA
1 d[A] 1
(8-9)
[A] dt TA
296 chapter 8 Dynamic and Fourier Transform NMR
In these expressions [A] is the initial concentration of protons at this site and
d[A]/dt is the rate at which this initial concentration "disappears." Thus, we are
concerned with the lifetime of the initial A and not with the bulk concentration
of A. What we measure is 1/-A and not the rate, so in a particular experiment a
first order decay constant is always observed.
Now consider the mechanism for the chemical reaction that is occurring
and causing A to leave site A. It could be a first-order process for which
d[A] = k[A] or k= - -dA- - 1 (8-10)

dt [A] dt TA
Therefore, for a first-order reaction, the rate constant k equals the 1/TA that is
measured. If several experiments are carried out at different concentrations of A
for a process that is first order in A, there will be a change in the rate; but there
will be no change in the observable from the nmr experiment, which is i/TA.
Thus, an observation of no change in lifetime with a change in concentration
corresponds to a first-order process.
Next consider a process that is second order in A:
d[A] 2
dt
or
1 d[A] k[A]=
[A] dt
Now in a series of experiments in which [A] is changed, 1/TA will also change
linearly with the concentration. Observation of this behavior suggests a second-
order reaction.
Next consider a reaction that is first order in A and first order in B:
d[A] = k[A][B]
dt
Now we must specify whether the A resonance or the B resonance is being studied.
If we are studying the A resonance, then we must use
1 d[A _ k[B] = /A, (8-12)

[A] dt
Now as the concentration of A is changed in a series of experiments, the value

of 1/TA remains constant; but as the concentration of B is changed in a set of
experiments, the value of 1/TA changes in direct proportion to the concentration
of B. The opposite result is obtained if the B resonance is examined.
In the system illustrated in Fig. 8-3, the exchange of free 2-picoline with the
four-coordinate complex, Co(2-pic) 2 Cl 2 , was studied 2 ) in deuterated acetone as
a function of the concentration of 2-picoline. The lifetime of the picoline on the
metal, r, was independent of the concentration of Co(2-pic) 2 Cl 2 , but varied
linearly with the free ligand concentration. A plot of 1/Tm vs. [ 2 -picoline] passed
through the origin. This indicates a reaction mechanism that is second order
overall: first order in Co(2-pic)2 Cl2 and first order in 2-picoline. The activation
parameters calculated in this article from a full line shape analysis were
comparedV with those from the line width equations. Errors in Eranging from
5 to 150% were introduced by the line width approach.
8-3 SOME APPLICATIONS OF NMR KINETIC STUDIES
Ligand exchange reactions have been studied for many transition metal ion
complexes. In many cases, the same compound serves as both the ligand and the
solvent. One cannot determine the order of the reaction in ligand when the ligand
and the solvent are the same, because its concentration cannot be varied.
Furthermore, if one wishes to study a series of ligands in a series of experiments,
one has varying contributions to the thermodynamic parameters obtained from
changing the solvent when the ligand is changed, since the same compound serves
both purposes. These problems are avoided, as in the system reported in Fig.
8-3, when the complex has solubility and can be studied in a non-coordinating
solvent.
In a study(6) of the exchange of free L (where L is [(CH 3)2 N] 3PO) with
CoL 2 Cl2 in CDCl 3 , the resulting kinetic data indicated that the reaction
proceeded by both first (i.e., dependent only on CoL 2 Cl2 ) and second (i.e., first
order in CoL 2 Cl2 and first order in L) order reaction paths. Since CDC13 is a
non-coordinating solvent, the first order path provides evidence for the existence,
of a three coordinate cobalt(II) complex as an intermediate.
The mechanisms for ligand substitution reactions in octahedral transition
metal ion complexes are difficult to ascertain. A recent study(7 a) of the exchange
of CH 30H with [Co(CH 3 OH), 2 +](BF 4 -)2 illustrates the problem. The value of
1/r, is independent of the metal complex or methanol concentration, supporting
a rate law:
Rate = nk [complex]
where n represents the coordination number of the exchanging ligand. The

following mechanisms(7 a) are consistent with this rate law and should be
considered:
Mechanism 1: SNI(lim) or D-type

k
1
Co(MeOH)6 2 , Co(MeOH),2+ + MeOH slow
k2
Co(MeOH), 2 + + MeOH* - Co(MeOH),(MeOH*) 2 + fast
for which the rate law is
Rate = kl[Co(MeOH)6 2
+]
Mechanism 2: SN(IP) or I,
Ko
Co(MeOH)6 2 + + MeOH . Co(MeOH)6 2+} {MeOH*} fast
{Co(MeOH) 62 +} {MeOH*} ±MCo(MeOH),(MeOH*)+MeOH

slow
which has the rate law
Rate = k 1KOS [Co]T[MeOH]

Kos[MeOH] + 1
where
[ColT = [Co(MeOH) 62 +] + [{Co(MeOH) 62+} {MeOH*}]
In the second mechanism, Kos is the outer sphere complex formation constant.
When Kos[CH30H] > 1, the rate law for mechanism 2 reduces to a form identical
to that for mechanism 1, and the two cannot be distinguished. If the [CH 30H]
could be reduced to a small value, then 1 > Kos[CH 3 OH] and second-order
kinetics would be observed, differentiating the two mechanisms. For a Kos of
about 1, changing the [CH 30H] from 3 M to 8 M would increase the rate
constant by only 10% and experimental error would make this change hard to
detect. A larger value of Kos would make the difference even smaller. At lower
[CH 30H], dissociation of some methanol from the complex, accompanied by
anion coordination, becomes a problem. These problems are described here
because they are common to many nmr studies on systems of this type. In order
to truly distinguish these types of mechanisms, it will be necessary to study
systems in which the free ligand concentration, [L], can be made sufficiently low
so that 1 > Kos[L].
In an nmr kinetic study( 7b) using a density matrix analysis the kinetic
parameters and the thermodynamics of adduct formation, show that the reaction
Ni(SDPT) - B + B' : Ni(SDPT) - B' + B
(where SDPT is a pentadentate ligand with two negative charges, B and B' is
4-methyl pyridine) proceeds by a pure dissociative mechanism involving a
five-coordinate, NiSDPT, intermediate. By comparing data obtained in toluene
and in CH 2 C12 as solvent, the role played by the latter solvent in the reaction
was clearly established.
The spectrum of N,N-dimethylacetamide, CH 3C(O)N(CH 3)2 , has three peaks
at room temperature, two of which correspond to the different environments of
the two methyl groups on the nitrogen (one cis and one trans to oxygen). The
C-N bond has multiple bond character and gives rise to an appreciable barrier
to rotation about this bond. As a result of this barrier, the two non-equivalent
N-CH 3 groups are detected. As the temperature is increased, the rate of rotation
about the C-N bond increases and the N-methyl resonances merge, giving rise
to a series of spectra similar to those in Fig. 8-1. The lifetime of a particular
configuration can be determined as a function of temperature, and the activation
energy for the barrier to rotation can be evaluated.(8') Similar studies have been
carried out on other amides(9) and on nitrosamines.(") The quadrupolar '4 N
nucleus makes a temperature dependent contribution to the line widths in some
of these systems, introducing error into the resulting parameters. The rates of
inversion of substituted 1,2-dithianes and 1,2-dioxanes("1 are among the many
other rates that have been studied.
The mechanism of proton exchange for solutions of methyl ammonium
chloride(1 2 ) in water as a function of pH was evaluated by an nmr procedure. In
acidic solution (pH = 1.0) the nmr spectra consist of a quadruplet methyl peak
(split by the three ammonium protons), a sharp water peak, and three broad
peaks from the ammonium protons. The triplet for the ammonium protons results
from nitrogen splitting. No fine structure is observed in the ammonium proton
peak from the expected coupling to the methyl protons. Partial quadrupole
relaxation by the nitrogen causes these peaks to be broader than 3 HNCH' As the
pH is increased, rapid proton exchange reactions begin to occur and the CH 3,
H 2 0, and NH 3 + bands begin to broaden. Eventually, at about pH 5, two peaks
with no fine structure remain, one from the protons on the CH 3 group and the
other a broad peak from an average of all other proton shifts. As the pH is raised
to 8, the broad proton peak sharpens again. The CH 3 broadening yields the
exchange rate of protons on the amine nitrogen; the broadening of the water
line measures the lifetime of the proton on water, and the broadening of the
NH 3 triplet measures the lifetime of the proton on the ammonium nitrogen.
An analysis of the kinetic data yields: (1) the rate law and a consistent mechanism
for the exchange
ki
CH 3 NH 3 ' + B : CH 3NH, + BH+
where the base B = H 20, CH 3NH 2, or OH -; (2) the fraction of the above
protolysis that involves water; and (3) the contribution to the exchange reaction
from:
CH 3 NH 2 + BH+ : CH 3 NH 3 + + B
and
CH 3 NH 2 + B : CH 3NH 2 + B'
(B' is B with one hydrogen replaced by a different hydrogen). The details of this
analysis can be obtained from the reference by Grunwald et al.," 2 which is highly
recommended reading. In subsequent studies the rates of proton exchange in
aqueous solutions containing NH 4 , (CH 3)2NH 2 *, and (CH 3 )3NH+ were
measured and compared." 3
Another example of information obtained from nmr rate studies is illus-
trated" 4 ) by the 9 F spectrum of SiF 2- in aqueous solution; Fig. 8-4. The main
peak arises from fluorines on 2 sSi (I = 0), and the two small peaks result from
spin-spin splitting of the fluorine by 2 9 Si (I '/2). The appearance of the satellites
corresponding to JSiF indicates that the rate of exchange of fluorine atoms must
be less than 103 sec-' [r' > 1/(v, - vB)]. It is also found that the spectrum of
solutions of SiF 6 2- containing added F- contains two separate fluorine reson-
ances. These are assigned to F- and SiF 6 2 -. There are satellites (JSi-F) on the
J29Si-F
FIGURE 8-4 The fluorine nmr spectrum

of SiF6 . See table inside back cover
(Properties of Selected Nuclei) for 2Si
natural abundance.
SiF 6 2 ~ peak. When a solution of SiF 6 2 is acidified, rapid exchange occurs, the
satellites disappear, and the central peak broadens. The following reactions are
proposed:
H2 0 HF
SiF 6 2- + H 3 0+ --. + HFSiF, -- : H 2OSiF, -- + SiF 6 2- + H 3 0+
-H 20 -HF
The 3 C spectrum of CO 2 in water gives rise to two peaks, 15 ) one from

dissolved CO 2 and a second from H2 CO 3, HCO3 -, and CO 3 2 -. Rapid proton
exchange gives rise to a single 3 C peak for these latter three species. The reaction
CO 2 + H 2 0 -* H2 CO 3 has a half-life of about 20 sec, so a separate peak for
dissolved CO 2 is detected.
8-4 INTRAMOLECULAR REARRANGEMENTS STUDIED BY

NMR-FLUXIONAL BEHAVIOR
The nmr spectrum of B3 H - is of interest because it demonstrates the effect of

intramolecular exchange on the nmr spectrum. The structure of B 3H,- is
illustrated in Fig. 8-5, along with a possible mechanism for the intramolecular
hydrogen exchange.(16-18) The "B spectrum is a nonet that results from a splitting
of three equivalent borons by eight equivalent protons. The process in Fig. 8-5
is very rapid, making the three borons equivalent and the eight protons equivalent
in the nmr spectrum. The eight hydrogen atoms remain attached to the boron
atoms of the same molecule during the kinetic process and the splitting does not
disappear. JBH is a time average of all the different B-H couplings in the molecule.
Contrast this to rapid intermolecular exchange, in which a single boron resonance
signal would result if exchange made all protons equivalent.
H H F H H -1H H
B B B
HH H H H" HH
H H H H H H H' H
FIGURE 8-5 The structure of B3 H and a proposed intermediate for the intra-
molecular exchange.
8-4 Intramolecular Rearrangements studied by NMR-Fluxional Behavior 301
+300
- 19 *
H
-250
H H -40-
-52*
H5 H, Fe- = -56'
H4 -600
H3 -64I
(A) _80'
(B) L
6.5 6.0.L. 5.5 . 4.4 4.0 I
3.5 3.0
FIGURE 8-6 The proton magnetic resonance spectra of (7r-C H )Fe(CO) CH. in CS
5 5 2 2
at various temperatures. The dotted line represents the resonance position of the
7r-C 5 H5 protons at each temperature. the amplitude of the +30 spectrum is shown
x 0.1 relative to the others. [Part (B) reprinted with permission from M. J. Bennett,
Jr., et al., J. Amer. Chem. Soc., 88, 4371 (1966). Copyright by the American
Chemical Society.]
Another possible explanation of the observed spectrum, that does not involve
exchange, is based on virtual coupling. Virtual coupling of the eight protons
would give rise to the observed nonet in the "B resonance. Strong boron coupling
has not been observed in the spectra of other boron hydrides.
Considerable information is available regarding the mechanism of exchange
processes from the non-symmetrical collapse of an nmr spectrum. An illustration
of this basic idea is provided(19) by the temperature dependence of the nmr
spectrum of (r/-C 5H,)Fe(CO)2 C 5H 5 , whose structure is shown in Fig. 8-6(A).
The spectrum obtained as a function of temperature is shown in Fig. 8-6(B).
The pi-bonded cyclopentadiene resonance gives rise to a single sharp peak at
3 = 4.4 at all temperatures because rapid rotation of the ring causes the five
protons to be equivalent. The remaining peaks in the spectrum arise from the
sigma-bonded cyclopentadiene ring. At -100 C, three distinct groups of reson-
ances are observed at - 6.3, - 6.0, and 3.5 ppm with relative intensities 2:2:1.
The most shielded resonance (5 = 3.5) is assigned to H 1. The H 2 and H5 protons
are isochronous, as are H3 and H 4 . The two sets differ only slightly in chemical
shift relative to the magnitude of their coupling constant and, as shown in a
subsequent section, this situation gives rise to complicated second-order spectra
that often contain more peaks than a simple analysis of the chemical shift
differences and spin-spin coupling would produce. The peaks at ~ 6.3 and - 6.0
are due to the 2, 3, 4, and 5 protons. The splitting from the H 1 proton is small;
in the limit where 3 JH1 -H 2 5 equals 4jH _H3 4 the two resonances at 6.3 and 6.0
would be mirror images. However, since JH1 -H 2 ,5 would be expected to be larger
than 4 JH -H 3 4 , additional fine structure would be expected in the peak assigned
to H 2 and H5 . With JHH 2 , 5 small, the additional fine structure would not be
FIGURE 8-7 Diagrams (1)
showing possible (5) (2)

intramolecular paths leading
to the nmr equivalence of the (4) (3)
a-CH 5 protons at room (A) (B) (C)
temperature. [Reprinted with
permission from M. J.
Bennett, Jr., et al., J. Amer. resolved but could be manifested in a broadening of the resonances. Using this
Chem. Soc., 88, 4371 (1966).
criterion, the deshielded multiplet with the broader resonance (6 ; 6.3) is assigned
Copyright by the American
to the 2 and 5 protons. Since these protons are isochronous, we shall label them
Chemical Society.]
as A and the 3,4 pair as B. As the temperature is raised, the A resonance collapses
faster than the B resonance. At -25 C, they are broadened and have completely
collapsed. At + 30 C, a sharp peak corresponding to a mole-fraction weighted
average resonance of the three types of protons is obtained.
Several mechanisms for this fluxional behavior are possible. The non-
symmetrical collapse of the spectrum rules out any bimolecular process (dis-
sociative or exchange in nature). A first-order dissociative process is ruled out
because the experimentally found activation enthalpy is too low. This low
activation enthalpy suggests some interaction between the pi orbitals of the ring
and the metal in the transition state. The three structures shown in Fig. 8-7 are
considered as possible transition states or intermediates and are referred to as
the r;5-cp (A), r 2-olefin (B) and rj3-allyl (C) mechanisms. The terms 1,2 shift and
1,3 shift have also been applied to the latter two mechanisms respectively.
The mechanism involving (A) can be eliminated for it would result in a
symmetrical collapse of the spectrum. The mechanisms involving (B) and (C)
would result in a non-symmetrical collapse, as shown by considering how the
labels in Fig. 8-7(B) are changed by a 1,2 shift. This is illustrated symbolically by
H A
A Ha
B : A
B B
-A B1
This analysis shows that all A's are changed into different types of protons, but
only half the B's are, so the A resonance should collapse more rapidly.
On the other hand, the 1,3 shift is described by
Ha
A
B
A
B > Ha
B A
LA J B
For this mechanism, the B resonance should collapse faster. Thus, to the extent
that the original assignment of the A and B resonances is correct, the 1,2 shift
is established. The intermediate or transition state in the 1,2 shift could be thought
of as one in which the it-electron density of the ring is arranged in a two-center
8-4 Intramolecular Rearrangements studied by NMR-Fluxional Behavior 303
71 bond coordinated to the iron, while the other three carbon atoms have an allyl
anion distribution of electron density.
Many fluxional systems have been discovered and are summarized in
reference 20. A procedure for a complete line shape analysis on systems of this
sort has been reported.""
There has been a very considerable effort devoted to nmr studies of the
intramolecular rearrangements of trigonal bipyramidal and octahedral complexes.
In 1951, Gutowsky and Hoffman 2 2 ) reported that the 19F nmr spectrum of PF,
was a doublet, even though electron diffraction had established the trigonal
bipyramidal structure of this molecule. Since P-F coupling is maintained, an
intramolecular process is required to equilibrate the fluorine atoms. Many other
systems show similar behavior e.g., Fe(CO), ("C nmr), CF 3 Co (CO)3 PF 3 , several
HM(PF3 )4 species, and HIr(CO) 2[P(C6 H) 31 2 . There have been several at-
tempts(2 3 ,2 4 ) to systematically enumerate all of the physically distinguishable
intramolecular modes for interconverting groups on a trigonal bipyramid. The
problem is a complex one, for it is necessary to insure that apparently different
pathways are physically distinguishable. The reader is referred to references 23
and 24 for details.
We shall briefly consider the fluxional behavior of CH 3Ir(COD)-
[CHP(CH 3)2] 2 to provide an illustration of the detailed mechanistic informa-
tion obtainable from work in this area of fluxionality.(2, 26') The static structure
of the molecule at low temperature is shown in Fig. 8-8(A). At -3 , the resonances
are assigned as in Fig. 8-8(C). The two methyls on each dimethylphenylphosphine
group are not equivalent (i.e., they are diastereotopic). Should the phosphorus
ligands change sites, they would become equivalent. The vinyl hydrogens, H,
and H 2 , are also non-equivalent. The phenyl resonances are not shown. As the
temperature is raised, the vinyl proton signal collapses into a singlet, but the
P-CH 3 proton resonance is not affected up to 67'.
The reasonable mechanistic paths that could exchange axial and equatorial
positions in the trigonal bipyramid are illustrated in Fig. 8-9. Scheme A involves
a twist of the diene about a pseudo-twofold axis through the metal and through
a point midway between the two double bonds. The intervening intermediate or
transition state is the distorted tetragonal pyramid labeled I.
Schemes B and C involve permutations of three sites. In B, one axial and
two equatorial ligands are interchanged by a rotation about a pseudo-threefold
axis constructed by drawing a line from the metal to the center of the face of the
trigonal bipyramid defined by double bond 1, double bond 2, and P 1 . This
rotation leads to the transition state labeled II. The process in C is best described
by considering the group R as being located in the center of the face of a
tetrahedron formed by P1, P 2 , double bond 1, and double bond 2. The R group
then moves through an edge and into the center of the face formed by P, P 2 ,
and double bond 1. The resulting structure has interchanged the two double
bonds. Scheme D proceeds through two trigonal bipyramid intermediate struc-
tures, Va and Vb, each of which involves a change of four ligands. Each change
occurs by a so-called Berry pseudorotation mechanism: two equatorial groups
open up their angle and the two axial groups move together in the direction in
which the equatorial angle is increasing, to form a distorted tetragonal pyramid.
The motion continues in this direction until the two axial bonds are equatorial
and the two equatorial bonds become axial. Structures Va and Vb are enantiomers,
and ready interconversion is expected. Paths in which the diene spans equatorial
CH'
CH CH3
CH 3 2
C6 H5 ,/
CH'\ 12
CHP
C6 H5 1
(A)
1170
870
FIGURE 8-8 (A) Structure, (B)

temperature dependent proton nmr 670
spectra, and (C) assignment of low
temperature spectrum of
CH3Ir(COD)[(P(C6 H5 )(CH 3)212. Solvent is
chlorobenzene and IMP refers to acetone
impurity present from recrystallization.
[Reprinted with permission from J. R.
Shapley and J. A. Osborn, Accts. Chem.
Res., 6, 305 (1973). Copyright by the 460
American Chemical Society.]
-3PMS
(B)
HH -CH J LJ-w
H2 H, -CH 2 CHE or CH ' CH 3

8-5 Spin Saturation Labeling 305
R R R
P P I
(A) 2 2
Pi 2
R R R FIGURE 8-9 Mechanistic

P2+; P schemes to account for
2 2P
(B) 1 axial-equatorial equilibration
1 of COD vinyl protons in the
2 complexes Rlr(COD)P 2,
where R is CH.. [Reprinted
with permission from J. R.
R P 2 2
I2 Shapley and J. A. Osborn,
(2 Accts. Chem. Res., 6, 305
(C) , 1 R 1
P3 (1973). Copyright by the
p1
R American Chemical Society.]
R 2 P1 R
2, 2 [P ] [R] 21 [P21
P2
(D) R R 1
sites have been eliminated as energetically unfavorable because of the geometrical

preference of this ligand for a 90 chelate angle.
Examination of these schemes indicate that B and C interchange the two
phosphorus ligands, but A and D do not. Thus, only the latter modes are consistent
with the observed spectral behavior. One cannot distinguish between the A and
D modes with nmr. There is a very extensive literature on this subject and, with
the example discussed here, we have indicated the kind of information that can
be obtained. For more details, the reader should consult references 25 and 26.
Studies involving fluxional behavior in six-coordinate complexes are described
in references 27-29.
8-5 SPIN SATURATION LABELING130 1
The double-resonance experiment can be used to label a proton and obtain

kinetic information. Consider molecules AD and A'D,
AD + A' -* A'D + A
where A and A' are the same molecules and the exchange rate is slow enough
to give two peaks in the nmr for A and AD. If we saturate a proton resonance
of A', this peak will disappear, but the exchange process will also cause a partial
saturation, i.e., a decrease in the intensity of the corresponding proton resonance
in AD, if the exchange rate of A' with AD is comparable to the relaxation rate
at the two sites. The lifetime for A, leaving a particular spin state in AD, TIAD,
then has contributions from TlAD and -AD (the lifetime of A at AD)
T1AD 1(TAD +TAD I) 1 (8-13)
If a saturating r.f. field is turned on at resonance A', saturation of this resonance

occurs immediately and one can observe (by sitting on the resonance of AD) an
asymptotic approach to a new equilibrium value of the magnetization at this
point. The plot of intensity versus time and the equilibrium value for the
magnetization can be analyzed for TIAD. When TIAD is known, TAD can be
calculated simply from the equilibrium value of the magnetization of AD.
When the field is turned off, the resonance at AD returns to its initial intensity
asymptotically, and this curve can also be analyzed to yield TAD. This procedure
is suited for the determination of reaction rates in the range between 10 3 and
1 sec - '. It thus can provide data on the slow side of the nmr line shape experiment
and is complementary to the line shape technique.
Spin saturation can also be thought of as equivalent to a deuterium labeling
experiment, and some of the same mechanistic information is available from this
technique as from the labeling experiment. For example, the heptamethylben-
zenonium ion is fluxional:
(1) (1)
CH 3
(2) 1- 3C CH 3 (2)
3C
1H3 CH3 H3C CH 3
CH 3
+ i- etc.
(3) (3)
H3C CH 3 H3C CH 3
(4) 11
M. (4 CH 3 CH 3
3 2
(a)
Does the interchange proceed by a 1,2 shift or by a random migration process?
At 28 , four fairly sharp resonances result, as shown in Fig. 8-10A. [There are
two types (1) CH, groups, two orthos, two metas, and one para.] Upon raising
(b) the temperature, the resonances broaden, etc., and a complete line shape analysis
suggests a 1,2 shift; but the differences between the expected spectrum for a 1,2
shift and that for the random process are subtle, and a simultaneous operation
of both processes was not ruled out. Saturation of the methyl resonance at site
(c)
2 produced spectrum (B) in Fig. 8-10, in which the intensity at site 1 is decreased.
No additional decrease in intensity is observed at site 1 when sites 2 and 3 are
saturated or when sites 2 and 4 are saturated, as shown in Fig. 8-10(C) and (D).
(d) If the random mechanism were operative, a further reduction in intensity at site
FIGURE 8-10 The spectrum I of 22% would be observed upon saturation at sites 3 and 4. The intensity
of the heptamethyl- doesn't change, within the 1 % accuracy of its determination, so the random
benzenonium ion; (B) process is minor or absent.
spectrum (A) with the (2)
proton decoupled; (C)
spectrum (A) with the (2) and 8-6 THE NUCLEAR OVERHAUSER EFFECT
(3) protons decoupled; (D)
spectrum (A) with the (2) and
An interesting phenomenon associated with the double resonance experiment is
(4) protons decoupled.
the Overhauser effect, discovered in the course of studying free radicals.( 3 1 a When
there is a coupling of the nuclear and electron spins, an enhancement in the
8-6 The Nuclear Overhauser Effect 307
intensity of the nmr transitions is observed when the esr transitions are saturated.
(N - 5
This same effect occurs in a nuclear-nuclear double-resonance experiment and is
called the nuclear Overhauser effect, nOe." 'I Changes in the intensity of a given
B
transition are observed when another transition is saturated or perturbed by the A
second r.f. field. Consider two uncoupled (J = 0) spin '/2 nuclei with the same y
but different chemical shifts. The energy level diagram in Fig. 8-11(A) results N r N
with populations indicated by N and 6. For a total of 4N nuclei, the populations (N -
2 ) ,B
'~2
A N + )
of each level are indicated as N with + J corresponding to the Boltzman excess

and -8 corresponding to the deficiency. The two transitions of A are degenerate (N+ 1
2
(J = 0) and give one line. The B transitions give another peak. The 22 -+ /# and
(A)
#o /x#transitions have Am = 2 and 0 and are forbidden. The population
differences for the transitions are 6 for the allowed A and B transitions, 0 for
Am 0 and 26 for Am = 2. Though transitions involving the latter processes - 1313
(Am 0 and 2) do not occur, they can provide relaxation mechanisms if the W W
Boltzmann distribution is disturbed. In Fig. 8-11(B), the assumed first-order rate W2
constants for the various paths are indicated by W. We will label these transitions
with the Am value as a subscript and the nucleus involved as a superscript. ap - ( Wa
If we were concerned with the T, relaxation of the A transition, it would w13--
depend on the rates for WA, W2 , and W. When W2 and W are zero we get our WB WA
simple definition of Ti:
Ti = 'W^ --aa
(B)
If W2 and W are non-zero in a multispin system, our T, experiment becomes FIGURE 8-11 (A) Transition
more complex. Now consider the case of an nOe experiment where we saturate in an A-B spin system
the A transition and observe B after saturation. After saturation the tAI#B -* 13A1B (N - B 1/2 and J = 0). The
levels of Fig. 8-11 have populations N - ('/2)6 while oActB - flA B populations equilibrium populations are
are N + ('/2)6. These levels are given in parentheses in Fig. 8-11. Thus the B given for each level and
transitions have population differences of 8, whereas the A is saturated and zero. those after saturation of the
Instead of 0 for Am = 0 and 28 for the Am = 2 transitions, their population A transition are given in
differences are now 8. We shall consider the adjustment back to the equilibrium parentheses. (B) Rate
system shown in Fig. 8-11. WA is of no concern for this is saturated. The processes that can contribute
to relaxation.
population difference across B is not affected for it is still 8. Thus, without W
and W2 there is no Overhauser effect. W transfers population from the #a to
the ocp state to get back to the zero population difference. This increases the
population of the excited state of one B transition and decreases the population
of the ground state of the other one decreasing the intensity of the B transition.
This is counterbalanced by W1B but if W is dominant, this will lead to a negative
nOe at B due to saturation at A.
The W2 process acts to transfer population to the 22 state to restore a
population difference of 28. This decreases an excited state and increases a ground
state population for the B transition. If it dominates WB, a positive nOe arises
at B from saturation at A. Solving the differential equations for the system leads to
(A) _W2 - W (8-14)

qB WB + W2 + W
for the Overhauser enhancement with

I - 1o
(8-15)
where I is the normal intensity and I is the intensity that is observed during
the perturbation. If nOe is observed, the W2 or W processes, referred to as cross
relaxation, must occur with the sign of the nOe indicating which one is dominant.
The cross-relaxation processes are dominated by dipolar coupling and W2
is the dominant term for small molecules in non-viscous solvents, with W
becoming dominant for large molecules or in viscous solution. These are the
most common cases, but there are other possibilities. Depending on the relative
signs and magnitudes of the g, values, there may be signal reduction instead of
enhancement, and even negative Overhauser enhancements (emission instead of
absorption). The enhancement has been treated quantitatively,(3 1") and it can be
shown that for the direct coupling mechanism the theoretical enhancement, q, is
given by:
(1±+71+ (8-16)
g2#2) 7'2
where the 1 subscript refers to the spin being saturated and the 2 subscript refers
to the spin being observed. The ideal enhancement is seldom observed because
of incomplete saturation and relaxation by other processes, e.g., T, relaxation.
If T, or 72 is negative, a negative enhancement results for the dipolar mechanism
leading in some cases to no peak (y1/72 = - 1) or to an inverted peak.
This effect can be used to indicate whether the coupling mechanism is direct
or indirect. Furthermore, by systematically observing the intensity changes in a
proton nmr spectrum as various protons are saturated, one can determine which
nuclei are in close proximity. The magnitude of the indirect coupling decreases
with the sixth power of the distance. Thus, one can distinguish cis-trans isomers
this way.
C,(CH 3 )
6
7_ Ph me
-N /,
Ru *'C~
PMe 3 H
6% C-H
Me3P
7 6 5 4 3 21
ppm
(B)
C-CH3
7 6 5 4 3 2 1
ppm
(A)
FIGURE 8-12 (A) The nmr spectrum of the complex indicated. (B) The nOe
difference spectrum with PMe 3 saturation.
8-7 Principles 309
An example of the nOe enhancement is shown(") in Fig. 8-12. The proton
nmr [(Figure 8-12(A)], with the C-CH3 doublet at ~1.5 ppm and the C-H
resonance at 4.75 does not distinguish between the structure indicated and one
in which the hydrogen and methyl of the chiral carbon are interchanged. The
spectrum in Fig. 8-12B is the difference spectrum between that in Fig. 8-12 and
one obtained when the methyl protons of the coordinated PMe 3 are irradiated.
With no nOe, a straight line would be observed and with nOe a peak
corresponding to enhancement results. The significant (6%) enhancement of the
C-H peak in the difference spectrum (indicated with an arrow) as well as the
disappearance of the C-CH3 doublet at 1.5 ppm indicates the compound has
the structure shown. Enhancements are also observed in the C6 (CH 3)6 and phenyl
resonances as would be expected for both isomers.
When pairs of free radicals are produced in solution, one may observe an
effect called chemically induced dynamic nuclear polarization (CIDNP) without
the need of saturating the electron spin transition to attain equal electron spin
state populations. The mechanism for the CIDNP process is involved, and the
reader is referred to reference 33 for more details.
8-7 PRINCIPLES Fourier

Transform
Many magnetic nuclei are present in nature in low abundance and also have low
sensitivity (e.g., "C). In examining the spectra of these materials, the tendency NMR
is to increase the r.f. power, but this often saturates the signal. Alternatively, one
can sweep the spectrum many times at low power and store the spectra in a
computer. This is called a CAT (computer averaged transients) experiment. The
noise is random and partially cancels (the signal is proportional to N and the
noise to N'12). The signal is reinforced by each sweep and, eventually, on adding
many sweeps, the spectrum emerges from the background. This process is time
consuming. Pulse techniques, referred to as Fourier transform (FT) nmr, are very
advantageous for this situation. In view of the many advantages of FTNMR to
be discussed in this chapter, this instrumentation now dominates the market.
In a magnetic field H, (z-direction), the equilibrium static magnetization is
shown in Fig. 8-13 as a bold arrow. The individual nuclei precess around the
z-axis with a Larmor frequency, o = yHN, where HN is the effective field felt at
the nucleus (HO(1 - a-) of equation 7-28). This leads to different frequencies of
rotation for each chemically shifted nucleus. The rotating xy-components of two
such nuclei, M, and M 2 , are shown in Fig. 8-13 where they are rotating with
Larmor frequencies o and w2. Now define a rotating coordinate system that
rotates in the xy-plane at the same angular frequency, w, as the Larmor frequency
of a reference compound e.g., TMS. If you were an observer on the rotating M, (<, t)
y-axis, for example, the TMS component and all nuclei with the same chemical
coil
shift (i.e., same w) would appear to be stationary. Nuclei in the sample that are
deshielded compared to that from TMS, precess at higher frequencies, wi, and FIGURE 8-13 Rotating
are moving slowly relative to the rotating frame at the rate w, - wo. Since the x,y-components, M1 and M2
nuclei are either standing still or moving slowly, this corresponds to an effective for two nuclei with different
field along the z-axis that is either zero or some small value. chemical shifts. MO is the
Since the effect of H0 has vanished, we can examine the effect of a weaker static magnetization aligned
applied magnetic field, H1, on the nuclei. In order to exert a constantforce, this along z field direction HO.
H 1 field has to be rotating at the frequency of the nuclei. This rotating magnetic
field is applied by a coil of wire around the x-axis energized by an oscillating
current at reference frequency w. The rotating H 1 field is a static vector in the
rotating frame and appears as such to the nuclei. When the H 1 field is on, it is
the only field seen by M0 and, as a result, MO begins to precess at the rate
o = yH, in the yz plane. For a "C nucleus, a 10-gauss rotating H 1 field would
produce one full rotation of M, (from +z to y to -z to -y back to z) in 100
microseconds. If the pulse is applied for 25 microseconds, the M, vector will
rotate 900 and point along the y axis. The angle 0, in radians, through which the
magnetization is tipped is given by:
0= yHit, (8-17)
where t, is the time of the pulse. The angle of tipping is called the flip angle and
the pulse causing a 90 flip is called a r/2 or 90 pulse. Doubling the t, of a 7r/2
pulse leads to a 180 or a 7r pulse.
In the nmr experiment, we monitor the y-component of M. After the H
1
field is shut off, the individual nuclei decay back to their H, precession. For a
nucleus rotating at w0 , our rotating frame and H 1 frequency, the y-component
decays as shown by the dashed line in Fig. 8-14. This is referred to as a free
induction decay curve and the y-component axis is labeled FID in Fig. 8-14.
These nuclei are rotating in phase with the rotating frame and the y-magnetization
slowly decays to zero as the z magnetization approaches its equilibrium value.
Now consider a nucleus that is rotating in the xy plane at a frequency o), different
than ao (M, of Fig. 8-13). These nuclei are moving faster or slower than the
rotating x,y frame and move in and out of phase with the y axis. The component
detected along the y axis during the time required for M, to decay to zero is
shown by the solid line in Fig. 8-14.
The FID curve is the superposition of four decay curves like that shown by
the solid line in Fig. 8-15. These four decay curves correspond to the four "C
resonances in the frequency spectrum that arise from the proton spin-spin
coupling. The FID "C spectra of molecules containing more nuclei are even
more complicated. 4 '4 4 That of progesterone, with the protons decoupled, is
given in Fig. 8-16.
My
FIGURE 8-14 A free

indication decay curve (A) 0 - - t
and its Fourier transform for
a single absorbing nucleus - (A)
(B). (The dashed line is
obtained when wo equals the
chemical shift, and the solid L_
line when o, is off resonance.) 200 100 0 -100 -200
Hz
(B)
8-7 Principles 311
1
-1 /J - ''
after
4+H +4transform
t
Fourier e
FIGURE 8-15 FID spectrum and its Fourier transform for 13CH31. (From T. C. Farrar
and E. D. Becker, "Introduction to Pulse and Fourier Transform NMR Methods,''
Academic Press, New York, 1971.)
In view of the complexity of this pattern, the FID interferogram is seldom

reported. When we are doing repeated scans, the FID curves are stored in the
computer, added together and Fourier transformed. Adding the individual
frequency spectra would be inefficient because of the computer time required to
do a Fourier transform. For a typical compound, we pulse for 10 - sec, measure
the free induction decay curve, store it in the computer, pulse again, measure the
FID curve, add it to the other one in the computer, and continue this for many
pulses. If the nucleus were sensitive enough, just one pulse would give an FID
curve that could be Fourier transformed to give the frequency spectrum. When
it is not sensitive enough, the entire spectrum is run many times and since one
FID curve is obtained in a few seconds, many spectra can be obtained in the
time required for a slow passage experiment. Basically, then, we measure the
decay of the magnetization in the u,v coordinates of the rotating frame in which
the net magnetizations of the different kinds of nuclei are precessing at their
Larmor frequencies.
FIGURE 8-16 '3C resonance of progresterone (proton noise decoupled). (From T.

C. Farrar and E. D. Becker. "Introduction to Pulse and Fourier Transform NMR
Methods," Academic Press, New York, 1971.)
FIGURE 8--17 A sequence of etc.

r.f. pulses.
10-6 sec 2 sec-j

time
The r.f. pulse is applied for a time t, (typical times are 10-' to 10-6 sec)
with a fixed frequency w and with very high power (100 watts). After waiting a
couple of seconds, the pulse is applied again, as shown in Fig. 8-17.
The pulse, if properly selected, causes all of the magnetic nuclei of a certain
element in the sample to absorb and eventually (after a transformation) leads to
a typical nmr spectrum. The problem now is to show how all these nuclei with
different Larmor frequencies can be made to undergo transitions by a pulse of
a single frequency wo. The mathematical function of time, f(t) in Fig. 8-17,
corresponding to a single pulse can be reproduced by summing together a series
of sine and cosine functions with different o's. This can be shown with a Fourier
transform. A Fourier transform corresponds to a generalized transformation in
function space. In this case, we wish to convert the time plot in Fig. 8-17
to the corresponding frequency plot, using equation (8-18):
F(w) = 21 L e'f(t) dt (8-18)
where f(t) is in terms of w, t,, and t. The result in Fig. 8-18 is obtained, where
A is the range of the principal frequency components. Consequently, pulsing our
frequency wo is comparable to having a whole distribution of frequencies available
to us from the r.f. source.
F(w)
it
I-
-j
c, --- (from e t
FIGURE 8-18 The frequency plot corresponding to the time plot of Fig. 8-17.
Probably the easiest way to see this is by considering how a square wave,
f(t), can be produced with a series of sine waves, i.e., a Fourier series. A square
wave is illustrated in Fig. 8-19.
8-7 Principles 313
In the limit as n - x, the following Fourier series will reproduce the square f(t)
k
wave function:
-7 0 t
f(t) = ao + Y (a. cos 2mnt + b. sin 2nnt) (8-19)

I
FIGURE 8-19 Graphical
Figure 8-20 shows how the superposition of the first three partial sums S, S 2 , representation of a square
and S3 corresponding to n = 1, 2, and 3, approaches a square wave. In Fig. wave.
8-20(A), we illustrate S,; in (B) the second term has been added to give S2 .
Adding the third term in (C) produces S3, which is beginning to resemble a square
wave. As n becomes very large, the resemblance becomes better.
In a similar way to that just described for a square wave, the distribution
of frequencies in Fig. 8-18 can be converted to A vs. t plots and added to give
the curve in Fig. 8-19. If there were no pulse, but just one continuous wave
approaching infinite time (the slow passage limit), only one frequency would be
required to describe this continuous wave, o),. As the time, t,, of a single pulse
decreases, the span of frequencies needed to describe this pulse increases. The
range of frequencies, A, in Fig. 8-18 is obtained from the Fourier transform of
the wave in Fig. 8-17 and is given by:
A = 47/t, (8-20)
r X
1 S2
FIGURE 8-20 Addition of
the first three waves in the
Fourier series leading to a
7 X square wave.
i X
where t, is the duration of the pulse. In the description of the single pulse, the
wave stops after time t,. A given frequency continues for infinite time. When the
pulse is passed through the sample, the appropriate frequencies in Fig. 8-18 are
absorbed by the sample, causing transitions. Therefore, the pulse must be short
enough to cover the distribution of expected spectral frequencies with similar
intensity frequency components. As we can see in Fig. 8-18, the intensity falls
off as one moves from we. Accordingly, t, should be much less than 47r/A to get
a reasonable distribution of similar intensity frequency components. For a typical
pulse of 10 psec, the flat central portion of Fig. 8-18, where the amplitude is
about 1 % of the peak value is about 16,000 Hz wide.
8-8 OPTIMIZING THE FTNMR EXPERIMENT
One of the major advantages to FTNMR is the ability to obtain spectra of dilute
solutions of sensitive nuclei or concentrated solutions of insensitive nuclei (i.e.,
nuclei with low magnetogyric ratios e.g., "N.) Several variables need to be
considered in order to optimize the experiment for difficult samples. These will
be considered briefly here to give an appreciation of the problem. The reader is
referred to references 35 and 36 for greater detail.
A single 90 pulse rotates the magnetization into the xy plane and produces
the maximum signal in the FID. If one then collects the FID for a period T equal
to 5T, 99.3% of the recovery back to the equilibrium z-component occurs. In
the usual case where T, > T2 , most of the instrument time is spent waiting instead
of collecting data. If the second pulse is followed before a r equal to 5 T1, a smaller
magnetization is observed. When r is 1.27T, only 72% of M, is collected, but
since more pulses can be employed in the same time, a 430% gain in sensitivity
results. If a small flip angle, e.g., 300, is used, M, is now 1/2 of M, (sin 30 ), but
since Mz is 85% (cos 30 ) of its original value, it doesn't take as long to decay
back to the equilibrium Mz and the second pulse can be applied faster. A gain
of 50% is achieved for r = T, and even higher when r < T1. The flip angle
approach is used to enhance signal to noise with the optimum flip angle given by:
cos O62 = e-T (8-21)
It is clear from this discussion that in a sample containing nuclei with varying
T's, the intensity distribution of the peaks can be distorted because of variations
in the amount of the FID curve obtained. It can also be appreciated that nuclei
with large T,'s can be completely missed, e.g., the carbons of metal carbonyls
are often not seen using the standard acquisition times because only a small
fraction of the FID curve is obtained.
Selection of the pulse frequency, w, is an additional consideration. Changing
w0 lets you vary the regions over which one would carry out a 500-Hz sweep,
for example. If o) is selected at a lower frequency than a peak in the spectrum,
this peak can appear as an echo on the other side of w. It often shows up as
an inverted peak with an irregular pulse.
Our final consideration is variation of the spectral width. As shown in
equation (8-20), shorter pulses give a wider range of frequencies enabling one to
increase the spectral width. However, going to a larger spectral width leads to a
decrease in digital resolution because of the limited number of storage channels
8-9 The Measurement of T by (FTNMR) 315
in the computer. Techniques such as zero filling, matched filters, and electronic
filters can also be used to improve resolution and enhance signal to noise. 3 6 )
8-9 THE MEASUREMENT OF T BY FTNMR
In a static field, the nuclear moments precess about the field direction as shown
in Fig. 8-21. As described earlier, when a secondary field H1 is applied, which
is in phase with the Larmor frequency, a torque is exerted that tends to make
the moment precess about H 1 . If we define a rotating coordinate system that
rotates as the Larmor frequency, we only have to worry about the torque from
Ho
H1 . If we consider the H1 direction to be perpendicular to the page, the cone is
so tipped that projection of the magnetic moment vectors in the xy plane gives FIGURE 8-21 Precession of
a net xy-component (see Fig. 8-22). Relaxation tends to restore the system to nuclear moments in a
the situation in Fig. 8-21. The torque is the cross product of the magnetic moment Zeeman experiment. (From
vector and H1 , so at resonance in the slow passage experiment, H, tends to tip T. C. Farra and E. D. Becker,
the net magnetic moment vector (which has no xy-component in the absence of "Introduction to Pulse and
Fourier Transform NMR
H1 ), inducing an xy-component. (When H 1 is applied, a and # are not eigenfunc-
Methods,'' Academic Press,
tions, but some linear combination of them is.) This net xy-magnetization is
New York, 1971.)
detected when one passes through resonance in the slow passage experiment.
In a pulse experiment, it is possible to tip the magnetization vector 90',
180 , or n , depending on the duration of the pulse. In all but the 180 pulse
experiments, an xy-component is induced. In a 1800 pulse experiment, we invert
the magnetic moment vector (180 inversion) from the position where H, = 0 torque
(i.e., from positive m. to negative m.) and do not generate any xy-component as
can be seen for the net moment in Fig. 8-23.
After the pulse is turned off and decay occurs, the magnitude of M. just
decreases at a rate governed by the longitudinal relaxation time T. The series
of arrows in Fig. 8-24 represent the decay of the M. vector with time. Since the
individual moments relax in a purely random manner, no xy-component results.
It is impossible to detect this decay of M. for it has no xy-component. Now
H, i to page
consider an experiment in which we hit the sample with a 180 pulse, followed
by a 90 monitoring pulse. We can then detect the magnetization. After waiting
Ho
for thermal equilibrium to be reestablished (usually a time corresponding to 5T1
is employed), we can again subject the sample to the same 180 pulse, but now FIGURE 8-22 Effect of a
wait a while, and then follow with a 90 monitoring pulse. Such an experiment secondary field H1.
is referred to by the symbolism 180-r-90. The process can be repeated by waiting
for a longer time r before applying the 90 monitoring pulse. The series of spectra
that result from a sample is shown in Fig. 8-25 (page 316), where the delay time
r, before the 90 pulse, increases from left to right.
Effect of FIGURE 8-23 The effect on

180* the magnetization of a 180
pulse pulse.
FIGURE 8-24 Decay of the

Mz component with time after 41
a180 pulse.
II
* 77w,
(A) (B) (C) (D) (E) (F) (G) (H)()
FIGURE 8-25 T1 determination of the 13C Resonances of allyl benzene

(C6H5CH2CH=CH2). (A) 0.469 sec; (B) 0.938 sec, (C) 1.875 sec, (D) 3.75 sec, (E) 7.5
sec, (F) 15 sec, (G) 30 sec, (H) 60 sec, (1) 120 sec after the 180o pulse. The T, value
for the CH2 and =CH2 resonances are 9.7 and 8.4 sec.
+
The peak assigned to the CH2 carbon bound to the phenyl ring is marked
with an asterisk. The magnetization of this carbon is seen to follow the pattern
shown in Fig. 8-24. If one plots the magnetization as a function of time, 'r, the
curve in Fig. 8-26 is obtained. This curve is of the form:
FIGURE 8-26 Change in M = M,(1 - 2e-4)

magnetization (peak height)
of the z-component with time where MO is the initial magnetization. The equation can be solved for T,.A value
following a 180 pulse. of 9.7 sec results for the carbon marked with the asterisk in Fig. 8-25. T, can
also be estimated at zero signal intensity from the relation t1, = T, In 2. NMR
instruments have the software to provide values of T, for all of the peaks in the
spectrum. Proceeding from right to left in Fig. 8-25, values of 9.7 sec for CH2;
68, 62, and 66 sec for the CDC13 triplet; 8.4, 7.4, 10.6, 10.6, 15.3, and 41.4 sec
result. The last value listed corresponds to the ring carbon to which the allyl
group is attached and is barely discernible in the spectra. The CDCl3 solvent T,
is not accurate because the delay time between experiments is not long enough.
An inverted resonance is obtained whenever the net magnetization of the
sample is opposed to the field, for the detection system senses this as an emission
H, H,
(transition from a state opposed to the field to one aligned), and vice versa for
(A) (B) net magnetization with the field. Another way of looking at this is that a 90'
FIGURE 8-27 The effect of a clockwise rotation about HI of a magnetic moment vector opposed to the field
90' pulse on (A) - Mz and gives rise to a different phase (180' different) than rotation by 90' of a vector
(B) -+ Mz (in the rotating aligned with the field. This is illustrated in Fig. 8-27, where1 is to be considered
frame). perpendicular to the page.
8-10 Use of T, for Peak Assignments 317
8-10 USE OF T1 FOR PEAK ASSIGNMENTS
We will digress for a moment from our discussion of nmr to make clear what is
meant by a correlation function. Correlation functions are employed in the
description of processes in which the value of x does not depend on t in a
completely definite way, i.e., a random time process. However, the average
dependence of x(t) on t can be written in terms of probability distributions. When
discussing the self-correlation of a variable, the function is referred to as an
autocorrelationfunction.
Correlation functions are useful for describing random processes e.g.,
Brownian motion of the fluctuating fields arising from the random motion of
molecules that cause relaxation. We can define a correlation function, R(r), for
a system in which the spin exchanges energy with some property (e.g., a magnetic
moment) that is in equilibrium with the lattice. When these motions correspond
to a large number, n, of independent and uncorrelated events, the time-dependent
correlation function is given by the relation:
T(t)= Kz ai(t)2 e ~*c (8-22)
where ai(t) is the magnitude of the property (e.g., the nuclear magnetic moment,
describing the n particles, with the summation over all particles i = 0 ... n. The
symbol t refers to time, r to a time increment and Tc a time constant whose
reciprocal is the molecular reorientation rate constant. If we consider a case in
which an assemblage of nuclear moments were all aligned, equation (8-22)
describes how the vectoral sum is randomized with time.
A plot of a correlation function for some molecular process that gives a
Lorentzian line in the frequency domain is given in Fig. 8-28. The correlation
time T, for a process in which R(r) is an exponential in the time domain is defined
as the time required to get through 1/e of the curve (curve is R(z) e -''). As we
shall see, the correlation time will be needed in order to interpret the values
obtained from T, measurements.
The following principles underlie the use of T, in making peak assignments
in the nmr spectrum of a complex molecule.
FIGURE 8-28 Correlation

function for a process
Rlr) yielding a Lorentzian line in
the frequency domain.
time
1. 13 C relaxation times of protonated carbons in large or asymmetric

molecules are dominated by dipolar interactions with the directly bonded protons.
The value of 1/T, for '3 C relaxation by a hydrogen comes from use of equation
(8-23) when molecular rotation is isotropic, the hydrogens are decoupled, and
(w)C + Hifteff < 1 (the so-called extreme narrowing limit).
1/T 1 = Nh 2 7C2 7H
2
rCH -6 Zff (8-23)
Here 70 and H are the gyromagnetic ratios of "C and 'H, N is the number of
directly attached hydrogens, rcH is the CH distance, and reff is the effective
correlation time for rotational reorientation. Assume that only directly bound
hydrogens contribute. Otherwise, we have to sum the rCH Teff terms for any other
proton that could contribute. In order for the limit (oc + wH}Teff < 1 to hold,
the atom must be rapidly rotating. The equation applies only under conditions
of proton decoupling where the scalar coupling is removed. However, under these
conditions, it should be emphasized that the magnetic moments of the protons
are not undergoing transitions fast enough to average the moment to zero in the
time required for a rotation, so dipolar relaxation still occurs. Under these
decoupled conditions, the Fermi contact coupling makes no contribution to the
relaxation mechanism and the dipolar effect dominates. It should be emphasized
that because of the inverse sixth power in rCH, dipolar contributions to 1/T, from
atoms a long distance away will be negligible. In proteins, long is usually
considered to be > 5 A but is longer in rigid molecules.
2. If two carbon atoms in a molecule have the same Te,,f, but one of them
is not protonated and the other is, then the non-protonated one will have a much
longer T than the protonated one.
3. Differences in T,,f for different carbons in the same molecule may arise
from anisotropic rotation of the molecule in solution and from the effects of
internal reorientation. For a more complete discussion of the effects of internal
rotation and of proton decoupling, the reader is referred to references 37 and 38.
Applications of these principles are illustrated in the T, measurements of
some of the "C resonances of cholesteryl chloride.( 39 ) The results are summarized
in Fig. 8-29. The protonated carbons on the ring backbone (not shown) and
other protonated carbons (also not shown) all have the same T and hence the
same Teff The overall reorientation of the molecule is isotropic, and T, values
can be used to distinguish CH and CH 2 protons. The carbon with no protons
attached, C(13), is seen to have a much larger T, (i.e., a sharper line when T, = T2 )
than others in the molecule. The methyl carbons have long T's, considering that
1.5 0.67 2.0

22 24 26
FIGURE 8-29 T1 values in 21
(sec for various "C atoms in 1.5 0 23 25
cholesteryl chloride in CC14 0.49 0.42 1.8
at 42 . [Reprinted with 0.51
permission from A. Allerhand 17 27
and D. Doddrell, J. Amer. 3 16 2.1
Chem. Soc., 93, 2777 (1971). 3.4 0.23
Copyright by the American 0.44 0.23
Chemical Society.] 14 15
8-12 13c 319
there are three protons on such a carbon. This is due to the rapid rotation about 26.9 86.4
the C-C bond that decreases zeff; as seen in equation (8-22) this decreases 1/T1 H H
or increases Ti. By comparing Ti's down the long side chain at the top of the
figure, one sees that the effect of internal motion increases toward the free end HO
of the chain as expected. 35.8
The proton T, values of vinyl acetate are shown in Fig. 8-30. The r 6 dipolar C
effects of the protons on each other are seen in all three values.
H3C 0
Advantage can be taken of the large differences in the T, values of protons
to simplify the nmr spectrum. For a particular delay time following the initial FIGURE 8-30 T, values for
180 pulse, the magnetization vector can decay to zero intensity. No intensity the protons of vinyl acetate.
will then be detected when this is followed by a monitoring 90 pulse (see Fig.
8-25 and the discussion of it). Thus, if there are two overlapping peaks with
different T's, a 180 pulse followed by a properly timed wait before the 90 pulse
will remove the peak with the shorter Ti. A threefold or greater differential in
the relaxation times of the overlapping protons is ideal for this application. With
T for HOD about 7 see and that of solutes usually less, this technique can be
used for solvent suppression.
8-11 NMR OF QUADRUPOLAR NUCLEI
The nmr lines of quadrupolar nuclei are very broad. For spherical rotation, the
width is a function of the nuclear quadrupole moment and the correlation time
ze as given in equation (8-24):
1 3 21±3 1 2
)\Iqe q 2
(8-24)
T, 40 P(21- 1) ( 3 h
Since the nuclear quadrupole effect on the relaxation process dominates the line
width and since this effect is intramolecular, only the rotational contribution to
T is important. However, it is possible to have both isotropic and anisotropic
rotation, and this complicates the line width interpretation. For example, in
CHCl 3, the rotational diffusion constant at room temperature perpendicular to
the C 3 axis, D,, is 0.96 x 10" sec-, whereas that parallel to this axis, D1 , is
1.8 x 10" sec'.
8-12 3
C Applications
and Strategies
A brief discussion of "C magnetic resonance, abbreviated cmr, will give us an
opportunity to review and apply several of the principles and phenomena we in FTNMR
have developed: the Fourier transform technique, the use of T data to assign
resonances, and the nuclear Overhauser enhancement that results in the 3C
spectrum by decoupling the proton nmr.
The total signal-to-noise gain when all of the techniques we have been
discussing are employed is interesting to note. In Fourier transform spectroscopy,
the multiplet collapse from double resonance, the nuclear Overhauser effect, the
larger size of sample tube, and the accumulation that can be done in the same
6Ts ppm
220 200 180 160 140 120 100 80 60 40 20 0

I I T I I I I IT I I I I | | I II
ALKANES - . --
CH3
-CH-
-CH
C --- - -
SUBSTITUTED CYCLOHEXANES (PENTANES)
UNSUBSTITUTED cCz n -4-17--

CHt-
(CHJ)- - - - - - -- - ---- --
ALCOHOLS
CH3OH
-CH2OH
CHOH
-C-OH
CHX,X F, CI, Br, I - . .
ALKYNES
=CH
C
ALKENES -- - - - -
=CH
-- CH
AROMATIC HYDROCARBONS
Monoarormatics
C-H
C-
Diarormatics
C-H
C-
C (bridge)
Triaromatics
C-H
C-
C (bridge)
HETEROAROMATICS - - -
Pyridines
ortho C-H
ortho C-
Thiophenes
ortho C--H
ortho C--
Furans
ortho C-H a
ortho C-
HETEROSUBSTITUTED AROMATICS - - - - -
C-OH
C-OR -
C-CHO I
C-COOR
C-COR
C-NH2
C-NR;
C-NO2
C-X, X = F, Cl, Br, I
ESTERS - - - -
-COOR
Conjugated
ACIDS - - - - -- - -
-COOH
Conjugated
ALDEHYDES - - -
CHO
Conjugated
KETONES - - -
Cyclic & Bicyclic

Alicyclic
Conjugated
-20 0 20 40 60 80 100 120 140 160 180 200 220

6
ces, ppm
FIGURE 8-31 A t3C correlation chart. [From R. K. Jensen and L. Petrakis, J. Mag.
Res., 7, 106 (1972).]
8-12 13c 321
time all lead to a total gain of ~ 104. Routine high resolution "C spectra of
organic molecules can be obtained on 0.5-M solutions in 20 minutes and on as
low a concentration as 0.05 M in 20 hours.
The cmr spectra are dominated by the local paramagnetic term leading to
a wider range of shifts than protons. The shifts observed provide a very good
indication of the functional groups present in a molecule, as can be seen from
the correlation chart in Fig. 8-30. The fact that a given carbon-containing
functional group (e.g., a substituted benzene or heterocycle containing ring
carbons) shows a very characteristic shift has made the fingerprint type of
application even more successful than in proton magnetic resonance.
Because of the nuclear Overhauser effect, the integrated intensity of a "C
resonance is not necessarily proportional to the number of carbons under the
peak. In some cases, this problem can be overcome by adding a free radical or
a paramagnetic ion to the solution. Spin-spin coupling from 3 C- 1 3 C is not seen
in the spectrum of a molecule containing several C-C bonds if it is not enriched
in 3 C because with a 1.10% natural abundance the probability that two 3C
nuclei would be next to each other in the molecule is very low, - 10-4.
In a large organic molecule, there are many "C resonances to be assigned.
The correlation chart in Fig. 8-31 will obviously not permit an assignment of
all the resonances. As mentioned earlier, T, measurements are a considerable aid.
There is another technique, referred to as off-center double resonance. As
mentioned earlier, if all the protons are decoupled in an nmr experiment, all the
information potentially available regarding the proton splitting is lost. If the
proton decoupling frequency employed is off center with respect to the proton
frequencies involved, only a partial collapse of the multiplet results and some
Overhauser enhancement is obtained. The resulting coupling constants and peak
intensities are distorted 4 01compared to those in the spectrum obtained without
any decoupling. Long-range coupling is eliminated. The partial multiplet structure
helps assign the resonances. This technique is more difficult at fields of 4.7 T or
more and has been replaced by polarization transfer and two-dimensional
techniques.
Since a major relaxation path involves the dipolar coupling of bound protons,
it is often found (particularly for non-protonated carbons, e.g., a metal carbonyl)
that a carbon signal can be saturated and is missing from the spectrum. To
observe nuclei with very long Ti's, small flip angles [0 of equation (8-15)] are
used in running the spectra and long pauses between pulses must be employed
to obtain the free induction decay curve. Even with these techniques, poor spectra
are often obtained. It has been shown that T, can be drastically reduced by
adding paramagnetic complexes, as mentioned earlier, for large fluctuating
magnetic fields arise from the paramagnetic complex moving through the solvent
and these lead to the very effective T, relaxation. Trisacetyl-acetonato chro-
mium(III) is a paramagnetic complex that is quite unreactive and has been
successfully employed this way.(4" A series of spectra in which the concentration
of the relaxation agent is varied should be studied to determine whether the
chemical shift of the "C resonance is being influenced by the paramagnetic
species. Doddrell and Allerhand(4 " have utilized a combination of these tech-
niques in the assignment of the resonances of vitamin B,,, coenzyme B,, and
other corrinoids. The reader is referred to the discussion in this reference for a
practical illustration of 1 3 C peak assignments.
Another interesting spectral enhancement technique is referred to as gated
decoupling.(43 ) When proton decoupling is terminated immediately before pulsing,

the "C-H coupling returns immediately, but the populations of the nuclear
energy levels are not equilibrated as rapidly. Thus, some Overhauser enhancement
remains. The technique thus involves decoupling, termination of decoupling,
pulsing, storage of the free induction decay, and repetition of this process for all
pulse cycles. Correct values of J"C-H are obtained.
Alternately, if the decoupling pulse is turned on just after the pulse, decoupling
occurs instantly but nOe develops slowly producing fully decoupled spectra with
little nOe distortion. One must wait for the nOe to decay fully between pulses.
This procedure has value in systems with negative nOe's
A very exciting application of cmr involves its use as a probe, permitting
one to employ stable isotopes as tracers.(44 ) Because the "C resonances are
extremely sensitive to the location of the atom in a molecule, the site of
incorporation of the 13 C can often be easily identified without the time-consuming
degradation required with radioactive isotopes. The incorporation of "C-labeled
glycine into coproporphyrin-III by the bacterium Rhodopseudomonasspheroidesis
an excellent illustration of the method. 44) The bacteria were grown on a medium
containing glycine that was 93 % labeled in the a-position with "C. The resulting
prophyrins had the "C atoms incorporated solely as pyrrole 21-carbon atoms
and methine bridge carbon atoms. This result is consistent with the sequence of
reactions proposed for the biosynthesis of porphyrins. If this type of tracer
application had been carried out using 4 C and subsequent radiochemical
analysis, the study would have been less definitive and more time consuming.
The product would have had to be degraded into small fragments to find the
location of the radioactive '4 C. There are many potential applications of this
type that involve using "C as a tracer atom. One of the main advantages to
using "C labels in biological systems is that this label does not disturb the
conformation of the biomolecule as a spin label or paramagnetic probe might.
In a "C study(45 ) of labeled CO binding to human hemoglobin, two separate
resonances could be observed for the coordinated CO. These occurred at 207.04
ppm and 206.60 ppm. The former resonance was assigned to CO bound to the
a-chain and the latter to that bound to the #-chain by studying an abnormal
hemoglobin that contains normal #-chains but 2-chains that do not bind CO.
In rabbit hemoglobin, three distinct iron(II)-binding sites were found. 4 6 ) A
functionally different hemoglobin subunit was established.
In another interesting application(47 ) of 13 C, T measurements were carried
out on selectively "C-labeled histidine bound to intracellular and extracellular
mouse hemoglobin. The intracellular and extracellular T values differed by only
25% suggesting that the viscosity of the intracellular fluid is not unusually large
(at least in this system). There had been considerable controversy regarding this
problem.
When molecules containing two directly bonded "C atoms are investigated,
spin-spin splitting results. The deviation of the intensity of the multiplet pattern
from statistical considerations in a biosynthetic experiment can provide a measure
of the correlation in the enrichment.(4 8 ) Such information has important mech-
anistic implications. For example, mixtures of doubly labeled and unlabeled
material can be studied. Dilution in the "IC- 3 C interaction would indicate
cleavage of this bond in the biosynthesis.
Cmr has been of considerable utility(4 9 ) in identifying the existence and
structure of carbonium ions and carbocations in solution. Typically, a positively
charged carbon will be deshielded compared to the analogous carbon in a neutral
reference compound, and it will cause an inductive deshielding of 5 to 15 ppm
in a neighbor atom.
The cmr spectra of many organometallic compounds have been reported.(50 54)
Complexation of olefins to AgNO 3 results(") in 1 to 4 ppm shifts, whereas shifts
of 30 to 110 ppm have been observed upon coordination to rhodium.!" The
benzene resonances in CHCr(CO) 3 are shielded(s2 ) some 30 ppm relative to
free benzene, whereas those of bound CO are usually deshielded from free Co.
The "CO resonances in diamagnetic metal carbonyls are relatively insensitive
to substituent and metal. Some typical results 5 0 ) are listed in Table 8-1. The
interpretation of the "C chemical shifts in organometallic compounds has recently
been criticized by Evans and Norton.( 5 ) A thorough analysis of the shift
interpretation and some results on the calculation of the shielding constant have
recently been summarized.( 6 ) The interpretation of small shift differences is very
difficult. Consistent interpretations based on increased deshielding with increased
metal-to-ligand 7E back-bonding have been offered.
TABLE 8-1 Some Typical "C Resonances in Metal

Carbonyls
Complex 8 "C of Carbonyl
Ni(CO)4 + 191.6
Fe(CO), +212
Cr(CO), +211.3
Mo(CO)6 +200.8
W(CO)6 +191.4
V(CO) 6 + 225.7
CpCr(CO)3 +246.8
CpMn(CO)3 +225.1
(CH 5 )3 PW(CO) +221.3 cis; +216.5 trans
(C6 H50) 3PW(CO) 5 + 217.6 cis; +213.9 trans
(Cp = cyclopentadienide)
8-13 OTHER NUCLEI
The advent of FTNMR enables the study of nuclei with low sensitivity(5 7) and
has led to a substantial increase in the number of elements readily studied. Various
nuclear properties of the elements are listed inside the front cover. The relative
receptivity of some of the more common nuclei are listed in Table 8-2. This
quantity indicates the relative signal strengths of solutions of equal concentration
and is proportional to y 3N(I + 1) where N is the natural abundance.
Quadrupolar nuclei in environments with cubic symmetry give reasonably
sharp lines that are broadened extensively in lower-symmetry environments. As
seen in Fig. 8-32, the sharp line for 3 Cl- in aqueous solution is broadened
extensively by an effect as subtle as ion pairing in R 3 NR''Cl- where R = n-octyl
and R' = CH3. The slower tumbling of the large ion pair leads to faster relaxation
and this coupled with the lower symmetry of chloride in the ion pair leads to a
broad line.
TABLE 8-2 NMR Properties of Selected Nuclei
Relative nmr Relative Relative nmr Relative

Isotope Spin Frequency/MHz Receptivity Isotope Spin Frequency/MHz Receptivity
107Ag 4.0 3.5 x 10- 14N

7.2 1.0 x 10
5
109Ag
27
4.7 4.9 x 10 2 31 N 10.1 3.9 x 10-
Al 26.1 2.1 x 10- Na 26.5 9.3 x 10
93
7sAs 17.2 2.5 x 10- 61Nb 24.5 4.9 x 10~
197Au 1.7 2.6 x 10- Ni 8.9 4.1 x 10-
10.7 3.9 x 10- 170
10B 13.6 1.1 x 10-
"B 32.1 1.3 x 10- 1870S
89 2.3 2.0 x 10-
11.1 7.9 x 10 1 0s 7.8
137Ba 31
3.9 x 10-
9
Be 14.1 1.4 x 10- P 40.5 6.6 x 10-
16.2 1.4 x 10- 207Pb 20.9 2.0
209B1i x 10-
05
79Br 25.1 4.0 x 10- '195
Pd 4.6 2.5 x 10-
81
Br 27.1 4.9 x 10- 87
Pt 21.4 3.4 x 10-
13C 25.1 1.8 x 10- Rb 32.8 4.9 x 10
43Ca 6.7 8.7 x 10- 185Re 22.7 5.1 x 10-
21.2 1.2 x 10- 187 Re
"'Cd 22.9 8.8 x 10
13Cd 22.2 1.3 x 10- 103Rh 3.2 3.2 x 10-
99
35C1 9.8 3.6 x 10- Ru 4.6 1.5 x 10-
37CI 8.2 6.7 x 10- 101Ru 5.2 2.8 x 10-
59Co 23.6 2.8 x 10 33S
7.7 1.7 x 10
53Cr 5.7 8.6 x 10- "2ISb 24.0 9.3 x 10
13.2 4.8 x 10- 123 Sb
'"Cs 13.0 2.0 x 10
63Cu 26.5 6.5 x 10- 45Sc 24.3 3.0 x 10-
65
CU 28.4 3.6 x 10- 77Se 19.1 5.3 x 10-
29
19F 94.1 8.3 x 10- Si 19.9 3.7 x 10
s'Fe 3.2 7.4 x 10~ "?Sn 35.6 3.5 x 10
69
Ga 24.0 4.2 x 10- 119Sn 37.3 4.5 x 10-
30.6 8
7'Ga 5.7 x 10- 7Sr 4.3 1.9 x 10-
t
73Ge 3.5 1.1 x 10- 18Ta 12.0 3.7 10 x
'H 100.0 1.00 "25Te 31.5 2.2 10-
x
2H 15.4 1.5 x 10- 47Ti 5.6 1.5 10-
x
177Hf 4.0 2.6 x 10- **Ti
203 5.6 2.1 10~
x
179Hf 2.5 7.4 x 10- 205
T1 57.1 5.7 10-
x
199Hg 17.9 9.8 x 10- T1 57.6 1.4 10-x
69
20'Hg 6.6 1.9 x 10 1 Tm 8.3 5.7 10-
x
1271 20.1 9.5 x 10- 51V 26.3 3.8 10-
x
22.0 3.4 x 10- i83W
"5In 4.2 1.1 x 10-
29
191lr 1.7 9.8 x 10 1 Xe 27.8 5.7 x 10
193 Ir 1.9 2.1 x 10- "1'Xe
89 8.2 5.9 x 10-
4.7 4.8 x 10- Y 4.9 1.2 x 10
13 9La 14.2 6.0 x 10- 17'Yb
67 17.6 7.8 x 10-
6Li 14.7 6.3 x 10- Zn 6.3 1.2 x 10-
9
7 Li 38.9 2.7 x 10 'Zr 9.3 1.1 x 10
25Mg 6.1 2.7 x 10-
ssMn 24.7 1.8 x 10-
95
Mo 6.5 5.1 x 10-
97
Mo 6.7 3.3 x 10-
Nitrogen nmr is an interesting example to consider( 58) for it illustrates the

advantages and difficulties associated with multinuclear magnetic resonance.
Compounds containing nitrogen exist in nine stable oxidation states and five
coordination numbers with lone pairs and nt-systems leading to interesting
perturbations. Two isotopes can be investigated "5N(I = '/2) and "4 N(I = 1). Both
have low sensitivity with receptivities of 3.9 x 10 6 and 1 x 10- respectively,
-,
compared to 3 C at 1.8 x 10-4. Higher fields help because sensitivity increases

with y73H 3 2. Relaxation is slow with "N slowing down spectral acquisition and
nOe is negative because 7 is negative leading to diminished signals with proton
decoupling. Enriched samples or large samples at high fields are required. 14 N
has a receptivity five times that of 13 C but is a quadrupolar nucleus. In addition
to environments with cubic symmetry, reasonably sharp lines are found in nitrates,
nitrites, nitro compounds, nitriles, and isonitriles. Correlation charts indicating
ranges of chemical shifts for a variety of nitrogen compounds can be found in
the literature.(s8 The range in chemical shifts found in nitrogen compounds is
about three times that found in 3 C. Deshielding of nitrogen is observed when
it is involved in nr-bonding to neighbor atoms and when it is bent as opposed to
linear in its coordination to metal complexes, e.g., bent vs. linear nitrosyl
coordination.
Deuterium nmr is of interest because its relaxation behavior is dominated
500 0 -500
by a quadrupolar mechanism and consequently is indicative of molecular
35
dynamics at the molecular position of substitution. In a simple case of rapid FIGURE 8-32 CI NMR of
isotropic motion, the longer T, the more mobile the C-D bond. The relationship (A) a 5-M aqueous sodium
becomes more complicated for more complex slower motion. Significant findings cloride; (B) ion-paired, 0.1 M
R4N+CI- in butanol, and (C)
involving the degree of molecular organization and dynamic processes occurring
1.2 M R4N*CI- in butanol.
in membranes have been reported.('9)
The advantages of multinuclear nmr in structural elucidation are nicely
illustrated by the nmr study(60) of:
Me Me Me Me
P-CH CH2--P
2
H-C Hg C
P -CH 2 CH2 - P
Me Me Me Me
The 1H, 3C, 3 'P, and i"Hg magnetic resonances can all be investigated.
The proton nmr shows the three signals expected for the CH 3 , CH 2 , and CH
protons with relative intensities of 12:4:1. Observation of one CH 3 and CH 2
resonance proves that the ligands are symmetrically bound to mercury. The CH 3
and CH 2 groups show couplings to 3P and satellites from the 16.8% abundant
199
Hg, whereas the C-H peak is a broad resonance with no resolution of
31
couplings. The proton decoupled P spectrum has a single main peak with
satellites from 'Hg and the proton-decoupled i"Hg resonance is a quintet
arising from coupling to four equivalent "P nuclei. No other structure is
consistent with all these observations.
More on 8-14 SPECTRAL DENSITY

Relaxation In equation (8-22), the autocorrelation function for a particle undergoing random
Processes motion was discussed. As mentioned earlier, the Fourier transform converts a
function from the time domain to the frequency domain. The Fourier transform
of the correlation function produces a quantity called the spectral density, J(w),
which is defined by
J(w) = R(r,) e'"c dr (8-25)
The spectral density for any given frequency, J(w), gives the intensity of the
fluctuation at that particular frequency. The fluctuations of importance in nmr
include the relative diffusion of one molecule past another, described by a
correlation time, TD; the rotation of a molecule about its rotation axes, R; a
rapidly relaxing nucleus, ts; and chemical exchange, T,. Classical texts on nmr
give equations for various contributions to relaxation, e.g., dipolar interaction,
quadrupolar interaction, chemical shift anisotropy, scalar coupling, spin-rotation,
and so on, in terms of the correlation function formalism. All of these phenomena
cause a given nucleus to experience a fluctuating field from another nucleus in
solution. The efficiency of nuclear relaxation will depend upon the intensity of
the frequency arising from the fluctuation, which in turn depends upon the
correlation times T, for the various processes. Accordingly, it is appropriate to
examine in more detail the relationship between spectral density and ze.
If we plot the spectral density vs. o, curves comparable to those shown in
Fig. 8-33 are obtained. These different curves correspond to different correlation
times, re, for our systems. (Recall that R depends upon the value of r.) The
half-intensity height indicated by the dots corresponds to the frequency w = T'
Therefore, only frequencies smaller than -r,-1 are available in the sample.
The flat portion of the curve corresponds to wro << 1, whereas the region to the
(A)
(B)
FIGURE 8-33 A plot of the
spectral density of various
frequencies for systems (A),
(B), and (C) corresponding to
different correlation times. (C)
I 1
106 108 10 l
w(sec- 1 )
8-14 Spectral Density 327
103
T1
E
-5 10-~1-
FIGURE 8-34 Dependences
of the T, and T2 relaxations.
10-5 -
\ T2
(C (B) (A)
10-9 1i i
1/5 10-11 10-- 10-3
Tc
right of the dot corresponds to wrec > L. The area under each of these curves is
the same. This is equivalent to saying that the kinetic energy in any two samples
at the same temperature is the same. This plot gives an intensity distribution for
any particular frequency; e.g., in the case of a magnetic nucleus undergoing
molecular motion, we have the intensity distribution for each frequency of the
oscillating fields arising from the complex motion that an assemblage of molecules
undergoes.
For the middle curve (B), a correlation time has been selected to maximize
J(w) at wo, (i.e., Ec-' is selected to be wo). Any curve with a larger or smaller Te
will have a smaller value of J(co) at o,. If we choose ow, at the Larmor frequency,
the magnitude of J(w9) at wo, will be proportional to the relaxation efficiency; i.e.,
the greater the intensity of the fluctuating moment, the more effective the
relaxation. Correlation times corresponding to curve (B) are most efficient, (C)
next most of those shown, and (A) the least efficient.
The T, relaxation is caused by frequencies that correspond to me,, whereas
T is caused by o- = 0 and we, frequency components. This is consistent with our
earlier description of the fluctuating field causing T, relaxation and the static
(co = 0) field in solids causing T2 relaxation. If we were to plot the behavior of
Ti as a function of re, we would obtain the curve labeled T, in Fig. 8-34. The
minimum corresponds to the solid dot of curve (B) in Fig. 8-33. Larger or smaller
correlation times give a smaller spectral density at this frequency. Curve (C) of
Fig. 8-33, corresponding to the dashed line labeled (C) in Fig. 8-34, gives the
behavior at shorter correlation times; a longer relaxation time results. It is
interesting to point out that T, is a double-valued function; i.e., a given T, can
correspond to two possible values of r,. To obtain ce, one must know on what
part of the curve the system is located.
The behavior of T is described by the curve so labeled in Fig. 8-34. Note
that for short correlation times, T parallels Ti; i.e., the spin-lattice relaxation
mechanism is randomizing the z-component and the xy-component equally. In
curve (A) of Fig. 8-33, we have a situation that would correspond to correlation
times in a viscous liquid or solid. There is a greatly reduced intensity of the
frequency component we, needed for longitudinal relaxation. However, the CO= 0
component is very large and the transverse relation is enhanced. The zero
frequency dipolar broadening thus decreases T2 , causing line broadening in

viscous liquids and solids. The increased T, associated with the increased
correlation time in this region of the curve explains why spectra of solids and
viscous liquid are easily saturated.
Often in the following pages we will note that processes that increase or
decrease the correlation time can sharpen nmr peaks, i.e., increase T, for the
protons. One should be able to deduce whether T, will increase or decrease by
knowing whether the spectral density change caused by the perturbation moves
the curve away from (B) toward (A) or (C). If the curve is broadened, the
perturbation moves the system from the (A) or (C) direction toward (B).
Multipulse 8-15 INTRODUCTION

Methods The multipulse methods all depend on variation in the start time, duration,
amplitudes, frequencies, and phases of a sequence of pulses. The measurement
of T, discussed earlier in Section 8-9 is a simple example of a multipulse sequence
to measure relaxation times. We can take advantage of pulse sequences to simplify
spectra, gain intensity, separate overlapping peaks to assign resonances, study
certain parts of the spectrum, suppress peaks, determine to which nuclei certain
nuclei are coupled even in heavily overlapped spectra, and determine which nuclei
are close by in space. The advances in this area have been phenomenal and
promise to continue. We shall briefly cover a few of the more common applications
here to illustrate the principles and the reader is referred to the literature 6 11 for
more details.
' Z'
(B) (C)
z z z
y'
X'D x *X)
(D) (E) (F)
FIGURE 8-35 The Hahn spin-echo experiment. (A) A 900 pulse applied along x' at
time 0 causes M to tip to the positive y' axis. (B) During the time T, the microscopic
moments in the xy-plane dephase because of field, H0 , inhomogeneity. (C) A 180*
pulse is applied along x'. (D) The moments regroup. (E) After a time Tfollowing the 180*
pulse, the moments rephase. (F) At subsequent times, the moments dephase again.
(From T. C. Farrar and E. D. Becker, "Introduction to Pulse and Fourier Transform
NMR Methods," Academic Press, New York, 1971.)
8-16 Spin Echoes 329
8-16 SPIN ECHOES
We can begin our introduction to this subject by considering a pulse sequence

that leads to a spin echo. A 90 pulse is applied along x' (Fig. 8-35) causing M
to tip to y'. After a time T, T 2 effects and field inhomogeneity at different places
in the sample tube lead to different rates of precession in the xy plane and cause
a fanning out of the magnetic moments in the xy plane. We shall refer to both
field homogeneity and other T2 effects as T2*, i.e., l/T 2* = 1/T2 + other effects.
Those nuclei precessing faster than average (the rotation rate of the rotating
frame) appear in the rotating frame to move toward the observer (i.e., looking
down the z' axis they appear to move clockwise), while those slower than average
move away [counterclockwise; Fig. 8-35(B)]. The shaded region in the figure
can be thought of as a whole series of vectors corresponding to nuclei precessing
at different frequencies. These are called isochromats. A 180' pulse is then applied
along x', moving all the moments 1800 about the x' axis [i.e., they are now fanned
out about -y' as shown in Fig. 8-35(C)]. It should be emphasized that a 180
rotation about x' is different from a 180 rotation about z. The former causes a
to become a' and b to become b'. While we are waiting another time period
corresponding to r, the faster nuclei move clockwise and the slower nuclei move
counterclockwise in the rotating frame, as shown by the arrows in Fig. 8-35(D).
However, now the slower nuclei move toward the observer and the faster ones
away from the observer, causing them to get back in phase. After the same time
r that was used after the 90 pulse, the isochromats have regrouped along y'.
The signal detected is shown in Fig. 8-36 and the peak at 2, is referred to as a
spin echo. If there had been no T relaxation by mechanisms other than field
inhomogeneity, we would have the same signal intensity in the echo as at the
start. We have, in effect, eliminated the field inhomogeneity contribution to T2
and can use the decrease in intensity to calculate T 2. The process -r-T can be
repeated from the peak of the echo to produce another echo. In practice, there
are two effects that reduce this amplitude: (1) T2 of the nucleus, which fans out
2 Tt X
ECHO
PEAK
FIGURE 8-36 The FID Signal for an Echo Sequence of Pulses. The signal is
recorded immediately after the nt/2 pulse and after a time, T, a-7r pulse is applied.
The signal rebuilds to a new peak at time 2z (i.e., t after the 7Zpulse) and then
decays. The process can be repeated by applying a 7n pulse at time 3T.
the magnetization randomly as opposed to the systematic nature of the applied

field, and (2) diffusion of molecules to different parts of the sample tube in the
time 2T. The latter introduces error in T2. We shall refer to the pulse sequence
(7/2) -- 2E-T-acquire as an echo pulse sequence.
Next, we consider how nuclei with different chemical shifts respond to the
echo pulse sequence. The 90 pulse rotates the magnetization of both nuclei to
the y-axis. In a rotating frame chosen to rotate at the frequency of the most
shielded a-nucleus, the other nucleus precesses faster leading after a time, -r, to
the situation shown in Fig. 8-37(C). The b-nucleus is precessing relative to a at
z Z Z z z
y ) yY
t =0 (n/2), T x,|
(A) (B3) (C (D) (E)
FIGURE 8-37 The effect of an echo pulse sequence on nuclei with two different
chemical shifts. This is a simplified representation of Figure 8-35.
a frequency in Hz corresponding to the chemical shift difference. After a 180'

pulse around x, b is behind a and catches up after an additional time T as shown
in Fig. 8-37(E). Thus, the chemical shifts are refocused at the peak of the echo.
Thus, we can carry out experiments on nuclei that are independent of field
inhomogeneity or chemical shift differences. We can now concern ourselves with
coupled systems without worrying about chemical shift differences.
First, let us consider the case of homonuclear coupled systems e.g., an A-X
I Y
proton-coupled system whose nmr spectrum consists of a doublet of doublets.

Focus on one of the doublets arising from coupling and let our frame rotate at
a frequency corresponding to the center of the doublet. Both components are
now precessing, one at +J/2 and the second at -J/2 Hz leading after a short
time to the situation in Fig. 8-38(C). Now when a 180' pulse is applied it converts
every a nucleus on A and X into a #and vice versa. All the nuclei precessing at
+J/2 with a neighbors now have #neighbors and precess at -J/2. The same
is true for the nuclei precessing at - J/2. Thus, the nr-pulse has changed the labels
as shown in Fig. 8-38(D). Now after a time r the isochromats are not refocused.
If we selected T-values corresponding to J/4, the isochromats would be aligned
along x in Fig. 8-38(E), and generate antiphase components at the echo. It is
Z Z Z Z Z
y ) y y y
(A) (BPC (D) (E)

FIGURE 8-38 The behavior of the A doublet in an A-X system to an echo pulse
sequence.
8-17 Sensitivity-Enhancement Methods 331
also important to point out that in a heteronuclear coupled system, the neighbor -AH-AC
nucleus is not inverted, the labels are not interchanged upon application of the
7n-pulse (i.e., in Fig. 8-38(D) above), and the components refocus in (E). With
appropriate equipment, pulses could be applied to both nuclei to exchange the -AH+AC
labels and lead to the case in Fig. 8-38. In all these experiments we should
HH,
remember that the magnetization is reappearing along the z axis with a time
H2
constant Ti.
,- AH-AC
8-17 SENSITIVITY-ENHANCEMENT METHODS

AH+AC
As can be seen by its receptivity (Table 8-2), the proton is the most sensitive
nucleus. When only small amounts of materials containing nuclei orders of
magnitude less sensitive are available, it becomes impossible to obtain their nmr
H 13 C I I
spectra by standard means. The nOe helps a little, but its advantages are limited.
Methods that transfer the favorable properties of the proton to other nuclei are
(B)
advantageous. Selective population inversion (SPI), insensitive nuclei enhanced
FIGURE 8-39 (A) Energy
by polarization transfer (INEPT), and distortionless enhanced polarization
level diagram including
transfer (DEPT) are methods of accomplishing polarization transfer. DEPT,
population differences for an
usually the method of choice, is more involved and requires an understanding AX system. (B)
of the principles of SPI and INEPT. We begin with SPI. Corresponding NMR spectra.
Consider two nuclei "C and 'H whose energy diagram is shown in Fig.
8-39. Coupling energies are so small on the energy scale of the diagram that they
are not seen. The magnetogyric ratio and Larmor frequency of hydrogen are
about four times those of carbon, so the transitions are four times as energetic.
With a total of N nuclei, each level would have a population of N/4 if they were
equally populated. We shall concentrate on the deviation from N/4, which for
the ground and excited states is represented by 2AH for the hydrogen transition
and 2AC for carbon; where AH equals 4AC from the magnetogyric ratios. If we AH-AC
were working with "C-H, the fourfold excess 'H population of the lower state
over that of carbon and the fourfold larger magnetic moment of the proton leads
-AH +AC
to a sixteenfold increase in the transverse magnetization when a n1/2 pulse is H, (inverted)
applied to the proton compared to carbon. The signal depends on the rate of
precession of the magnetization, which is four times greater for the proton than
H2
"C thus a 64-fold difference results. The 1" 3 Rh nucleus has a y that is 30% that
-AH -AC
of the proton leading to a sensitivity 32,000 times less. In general, the relative
signals are proportional to y and acquisition times to 7'.
AH+AC
Now consider the effect of inverting one line of the proton doublet H,.
This inverts the H, populations leading to the situation in Fig. 8-40. The (A)
population differences across the proton transitions are the same except one
is negative. However, the carbon transitions that previously were 2AC 13
are now (AH + AC) - (-AH - AC) or 2AH + 2AC for C, and HT c
(-AH + AC) - (AH - AC) or -2AH + 2AC for C 2 . We have transferred the
proton population differences to the carbon and added them to the existing (B)
differences. With AH = 4AC, we get 10AC from 2AH + 2AC or five times the
original intensity (which was 2AC) and we get 6AC from -2AH + 2AC or minus FIGURE 8-40 (A) Population
three times the original intensity. This leads to the "C spectrum shown in Fig. differences for an AX system
8-40(B). after saturation of H, (one
It would be tedious to analyze a complex spectrum this way for we would component of the doublet).
have to find and selectively invert one line of each proton doublet. Furthermore, (B) The corresponding nmr
we cannot obtain a "C spectrum that is proton decoupled. The INEPT spectra.
') y y ' y y y
x x x x x
1 n
H -
Acquire
FIGURE 8-41 INEPT pulse sequence.
experiment puts pairs of proton transitions in antiphase independent of their

chemical shift by using a non-selective pulse. This is accomplished by using a
spin echo from the pulse sequence shown in Fig. 8-41.
After a (7r/2), pulse to the protons, we wait for a time of J/4 so doublet
components from spin-spin splitting move through 1/8 of the cycle before the
it-pulse is applied (the components precess at ±J/2 Hz). The sensitive nucleus,
S, (in this case the proton) is rotated into the other half of the xy-plane. A
simultaneous application of the it-pulse to the insensitive nucleus (in this case
the carbon) causes a reversal in the sense of the precession of the J components
causing them to move as shown in Fig. 8-41. The chemical shifts and field
inhomogeneity will be refocused in the next J/4, but the magnetization due to
spin coupling is not, and ends up on the x-axis. A 90 pulse to the protons gives
us the antiphase components and the population enhancements we had in SPI
for all the protons. The simultaneous 90 pulse to "C enables us to acquire the
"C signal. Each "C spin-spin doublet gives rise to a positive and negative peak.
We have transferred the sensitive nucleus population difference to the insensitive
nucleus, but the positive and negative peaks have unequal intensities, e.g., +5
and -3 for a "C doublet. Various pulse sequences enable us to eliminate this
and to decouple proton splittings. The DEPT sequence has replaced INEPT and
gives rise to a spectrum with the appearance of a normal nmr spectrum.
A further advantage of these polarization enhancement experiments arises
from the fact that only the population difference of the sensitive nucleus, e.g., the
protons, leads to the signal intensity, so it is the T of this nucleus that determines
the pulse repetition rate. This is a big advantage for the study of insensitive nuclei
with long T1's (e.g., "N and 2 9Si). A further advantage arises in the application
of polarization enhancement techniques to nuclei that have negative nOe
enhancements. For nuclei that do not have large proton coupling constants (e.g.,
some rhodium-phosphine complexes), other sensitive nuclei (e.g., phosphorus)
can be used if a probe tuned to the two frequencies and two broadband
transmitters are available.
8-18 SELECTIVE EXCITATION AND SUPPRESSION
The DEPT pulse sequence can be used to select certain multiplets for study.
Pulse sequences can be used to selectively study all the CH or CH 2 or CH 3
8-18 Selective Excitation and Suppression 333
resonances; this is called spectrum editing. As an illustration of this application,
consider the "C nmr of
C6 H5 CH
CH CH2
Pd
H 3C 1CH3
N N
H3C / CH3
CH2 -CH 2
which contains thirteen "C resonances. Assignment is difficult. The chemical

shifts of CH 3 , CH 2 , and CH protons can be obtained(6 2 ) by selectively studying
each group of resonances as shown in Figure 8-42. All the coupling constants
FIGURE 8-42 The edited 1C
spectra of
Pd(<CHCHCH 2)(CH) 2NCH 2-
CH2N(CH 3)2 using the DEPT
pulse sequence. (A) The CH3
groups; (B) the CH2
(B) resonances; and (C) the CH
resonances. This is a
different pulse sequence
than INEPT. The spectra
come from mathematical
operations on several
spectra with different pulse
lengths.
(C)
140 100 60
(D)
in each group should be within about 30 % of their average value. The advantages
illustrated are even more important in analyzing even more complicated spectra.
The digitizer (- 12 bit) that converts the spectrum to digital form has a
dynamic range of about 2000 to 1. If the largest signal in the spectrum is just
equal to the capacity of the analogue-to-digital converter, signals 2000 times
smaller than this will not be detected at all. No amount of time averaging will
help. Because of the digitization process, a peak outside the spectral width will
fold back into the spectrum. Peak suppression or tailored excitation can alleviate
the problem. Peak suppression is best accomplished by pre-saturation if the other
needed components of the spectrum are not undergoing exchange with the
saturated peak. The irradiation of the resonance that induces saturation is turned
off immediately before the application of the pulses and acquisition of the
spectrum. This procedure can be used in conjunction with some of the pulse
sequences discussed so far and with the 1331 experiment to be discussed shortly
[1 refers to ('/ 4)m, 3 to (3/ 4 )nr and the line over the number indicates a negative
pulse, e.g., along the -x-axis].
Tailored excitations attempt to tailor the frequency distribution of the r.f.
excitation to avoid frequencies corresponding to certain regions of the spectrum
(recall the distribution shown in Fig. 8-18). The sequence, (ir/2)x-z-(r/2), acquire,
known as jump and return (JR), illustrates the principles for several experiments
of this type. If the transmitter frequency is set at the peak to be suppressed, it
will remain static in the T-interval while others will precess. The second (r/2)_
pulse returns the peak to be suppressed to the z-axis with no x-component
generated, i.e., no peak. The peak of interest moves relative to the rotating frame
during T, thus an x-component remains after the second pulse. This is illustrated
in Fig. 8-43 where A is the resonance to be suppressed and B the peak of interest.
Z Z Z
A-
FIGURE 8-43 The effect of A
the JR pulse sequence on a Y Y y
suppressed resonance A A and A
and one of interest B. x x x
(x/21 - (n/2)_
Ideally, T should be selected so the B magnetization lies along X. Other pulse

sequences have been utilized that are more effective than JR. The (n1/4),-
r-(3nr/4)_,-x-(3nr/4),-T-(nz/4)_, sequence, referred to as a 1131 sequence, is particu-
larly effective with r selected to optimize the spectral region of interest. This
sequence can provide a 1000-fold better reduction in a solvent resonance than
pre-saturation. For nuclei other than protons, the DANTE sequence can be
utilized for selective excitation. This consists of a train of hard pulses of very
small flip angles with constant phase and constant separation.
8-19 TWO-DIMENSIONAL NMR
The two-dimensional, (2-D) nmr experiment involves the following sequence of

events:
PREPARATION - PULSE t2
T Evolution time Detection time
The detection period corresponds to the usual collection of the FID, whose
transform leads to the usual frequency spectrum. To examine the rest of the
z z z z
MI I FIGURE 8-44 Behavior of
/ Y,y the magnetization during a
Y )_+
Y Zs -- Y
l~
'I
/
y _ _ (r/2), - t, - (7/2). sequence.
The x- and y-components are
x x x M = Moe / Ix shown in (C) and only the
rotation of the components
(A) (B) (C) (D)
by the 7r/2 pulse is illustrated
Preparation (n/2), t, (nt/2), - Detection in (D).
FIGURE 8-45 Nuclear

magnetic resonance spectra
detected at increasing t, by
increments of At1 for the
pulse scheme in Fig. 8-44.
V2
sequence and some important consequences, we shall consider a very simple

experiment involving the proton nmr of chloroform, CHC13 . Our preparation
time will consist of a (7E/2), pulse. We will wait for a time t,, after which time
another (nr/2). pulse is carried out and the signal detected. The behavior of the
magnetization during this pulse sequence is shown in Fig. 8-44. We shall ignore
T, relaxation, for simplicity, during the time t, (which is not related to T) and
concern ourselves only with the x,y-magnetization in the rotating frame. Fig.
8-44C shows the precession of the magnetization at v Hz forming an angle of
2nvt during the t, interval. The M, and M. components are shown and their
rotation by a (it/2). pulse is shown in (d). The M, component, given by M
sin 2nrvt 1 , is detected, Fourier transformed, and a frequency spectrum is obtained.
FIGURE 8-46 The
Now consider a series of n experiments in which t, is increased by an increment interferrogram obtained by
At, each time. The M, magnetization and resulting spectral intensity will vary plotting the amplitude of a
as t, is varied leading to the series of 17 spectra shown in Fig. 8-45 for 17 of slice parallel to the t1-axis of
our n experiments with regular At, increments. Fig. 8-45.
Now consider a slice through Fig. 8-45 at fixed v2 corresponding to the top
of the CHCl, peak. Plotting the variation in amplitude vs. t,, for not just the 17
slices but, for all n of the slices of our original experiment, gives Fig. 8-46. The
amplitude is oscillating with a frequency v corresponding to the chemical shift
of CHC13 because M. is given by Msin 2nvt,. It decays exponentially with a
time constant T2 . If we Fourier transform the curve in Fig. 8-46 we obtain the
frequency spectrum in Fig. 8-47(A), with its peak at a frequency corresponding
to v (the chemical shift of CHCl 3) because the transverse magnetization is
oscillating sinusoidally at Msin 2ntvti. If we now take a whole series of slices
through Fig. 8-45 at different values of v2 and stack these spectra together, we
get the plot shown in Fig. 8-47(B). The stacked plot is often presented as a
contour plot as shown in Fig. 8-47(C). Generally two-dimensional (2-D) plots
contain v2 along the abscissa and vi along the ordinate and this change is made
in Fig. 8-47. The symbols F 1 and F 2 are often employed to indicate the result
after Fourier transform of t, and t 2 to the frequency domain.
(A)
FIGURE 8-47 (A) The (B)

Fourier transfer of Fig. 8-46.
(B) The stacked plot of
Fourier transforms of various
slices of Fig. 8-46. (C)
Contour plot of Fig. 8-47B.
v1
6
V2
(C)
We have created a two-dimensional frequency spectrum f(vi, v2 ) by Fourier

transforming the two time dependencies ti, t 2 . In this example, we get a square
spectrum with both axes representing chemical shifts and a peak on the diagonal
in the frequency domain at v, v. We have not learned any new information from
the simple example chosen but have selected it to illustrate the 2-D method. Now
we move to more interesting 2-D systems and will begin by considering
homonuclear 2-D nmr. In principle, we can perform a 2-D experiment on any
system in which a modulation of the frequency spectrum occurs during the
evolution time, t. In the above example, the same modulation occurred during
both t, and t2 . Significant results are obtained when different processes occur
during the two time intervals. Generally, normal chemical shifts and couplings
correspond to t 2 , and.when a different modulation occurs on t1 , off-diagonal or
cross peaks occur in the plot of v, vs. v2 . To interpret the resulting contour plot
we need to know what the two axes represent and how the two magnetizations
are related. In all 2-D experiments, the F2 dimension contains information about
the precession frequency of the observed nucleus. The F1 dimension contains
information about the interaction of the observed nucleus and other magnetic
nuclei that develops during the evolution period.
Consider what happens during the second pulse for a case in which there
is a homonuclear coupling. To simplify the discussion, consider an experiment
in which one automatically increments the decoupling frequency across the region
of interest. Each decoupling frequency would manifest itself by a change at coupled
protons when their partner is irradiated. It would be time consuming to march
across all the frequencies. Instead, we will do this experiment by using FT in the
coupling dimension. This pulse sequence ((r/2),-t,-(nr/2),) is referred to as Jeener's
or the COSY experiment. In a homonuclear, AX, system the second (7r/2), pulse
causes magnetization to be redistributed among all the transitions with which a
given transition is associated. In an A-X system (four peaks), a given A transition 6.8 6.3
is associated with the other A peak and the two X peaks and likewise for the ppm
other A and two X transitions. Thus a given line in t2 may have components of FIGURE 8-48 The contour
its amplitude modulated as functions of the frequencies of all the other lines in plot of a homonuclear
ti. When the two Fourier transforms are carried out, this redistribution gives two-dimensional A-X system
rise to cross peaks in the 2-D spectrum. The resulting contour plot for an AX (/ = 1/2 for A and X).
system is shown in Fig. 8-48. The peaks along the diagonal arise from
magnetization components that have the same frequency during ti and t 2 (recall
CHCl3 ).This is the portion of the magnetization that was not transferred elsewhere
during the second pulse. Thus, looking along the diagonal, we see the normal
spectrum-a doublet of doublets. The cross peaks corresponding to the small
cluster arise from redistribution within a given multiplet (A1 A 2 or X1X 2 ). The
off-diagonal peaks farther off the diagonal correspond to redistribution between
different nuclei. In a more complex molecule we can determine which protons
are coupled to each other by the existence of off-diagonal peaks. We can think
of the coupling as providing a pathway through which the magnetization can
travel during the second pulse. Couplings that are less than the natural line width
and hence unresolved in the one-dimensional experiment, are easily detected in
COSY.
COSY has had important applications in cluster chemistry. Typical examples
include "B clusters(6 3) and .83 W clusters.( 64 ) For quadrupolar nuclei, the
relaxation times must be long enough that the coupling is not completely
eliminated.
The INADEQUATE pulse sequence enables one to detect natural abundance
3
C- 3 C couplings. In this experiment, the peaks from lone 3 C nuclei are
eliminated and only those involved in coupling remain. At each existing frequency
in the v, spectrum, a row exists parallel to v2 corresponding to the coupling of
adjacent "C atoms.
In the COSY experiment we have ignored the transverse magnetization
returned to the z-axis by the second pulse. This magnetization is also modulated
during t,. Since the component is along the z-axis, it is not detected. If processes
of interest modulated this z-component, a 900 pulse could be applied and its
behavior studied. Two such effects are chemical exchange and nOe. If we wait
2
a while Tm after the second pulse, i.e., the sequence /2-ti-7/2T-r/ -t2, a nucleus
whose z magnetization was modulated by one chemical shift during t, migrates
to another site during Tm. Thus, it has a different chemical shift during t 2 and
the resulting 2-D spectrum will have cross peaks between the peaks of exchanging
sites. Rate data is not obtained, but the peaks undergoing exchange are identified.
In a similar fashion pairs of nuclei, that show nOe in a one-dimensional
experiment, show cross peaks in this two-dimensional experiment. Spatial
CH3 relationships between protons can be determined by knowing which ones undergo
- CH 3 nOe. This technique is particularly useful in large molecules where nOe's are large.
The heteronuclear correlated 2-D experiment, is similar in concept to INEPT.
A 7r/2-t, period on the sensitive, S, nucleus is followed by a 7r/2 pulse on both
CH
2 F,(3)the sensitive and insensitive, I, nucleus followed by accumulation of the signal
F('C of I during t 2 . This is referred to as the heteronuclear shift correlation (HSC)
sequence. The 2-D plot gives V2 (S) VS. vl(I) i.e., a peak appears in the contour
plot for correlated nuclei at the chemical shift of I on the V2 axis and S oni the
-CH 2 OH v2 axis. In order to remove coupling, a it-pulse is applied to the I nucleus halfway
C HOH through t, The advantage of HSC is in assigning coupling relationships 3 in
heteronuclear systems. If the proton spectra is assigned, assignment of the ' C
F2 (H spectra in the 2-D plot is trivial. In weakly coupled systems, contour peaks appear
FIGURE 8-49 The only between protons and their directly bonded heteronuclear partners. In the
correlated heteronuclear presence of strong coupling between HA and H, in HA_-C-C-HB, cross peaks
('H, 130) two-dimensional to the non-bonded carbon can appear. Often the "3 C spectra or that of some
spectrum of butane-i ,3-diol. nucleus other than H is easier to assign than hydrogen because of less homonuclear
coupling in the former. These ideas are illustrated by the correlated, heteronuclear
(I'H, 13 CQ 2-D contour plot of butane-1,3-diol in Fig. 8-49. The '3 C peaks are
well separated and their assignment leads to assignment of the directly bonded
proton peaks. The 2-D spectra can then be used to assign the hydrogen spectra.
Taking a slice through the HSC spectra parallel to v, at the chemical shift of a
nucleus yields the proton spectra of its attached protons.(")~ This method is
particularly useful for locating proton peaks that are obscured by an unrelated
multiplet whose attached 13C resonances are well separated. Combination of the
COSY experiment (homonuclear correlated) on the protons with the HSC spectra
often allows the complete framework of a molecule to be traced.
Our final topic in 2-D nmr will involve J-spectroscopy. This involves spin
echo pulse sequences on the nucleus, observed during t2, to provide information
about coupling. Recall our discussion in Section 8-16 where we showed the echo
could be used to cancel out chemical shift differences to let us study couplings.
The proton decoupled spectrum will be along one axis and the carbon-proton
or proton-proton coupling constants spread out along the other. The decoupling
is modulated during the evolution period of the echo sequence with gated
decoupling or other techniques. The technique aids in the assignment of the
resonances and peaks due to spin-spin coupling in badly overlapping spectra.
The proton decoupled "3 C spectrum of a mixture of deuterated derivatives of
toluene is shown in Fig. 8-50(A). (Deuterium has an 1-value of one and C-D
couplings lead to the spectral complexity.) The J-resolved 2-D nmr is run by
employing gated deuterium (2 H) decoupling producing the 13 C- 2 H coupling
information on the F, axis. The "3 C chemical shifts of the four carbons are seen
on the F, axis. The slices parallel toF, and through the different carbons show
the single, three, five, and seven peaks for coupling constants of the substituents
CH 3, CHD, CHD2, and CD,. A stick diagram of overlapping resonances can
now be readily constructed to interpret the spectra in Fig. 8-50(A).
In spectra that contain both homonuclear and heteronuclear coupling, J
spectroscopy can be used to separate and assign the homonuclear and hetero-
nuclear coupling constants, e.g., proton-proton couplings can be distinguished
F2(Hz)
+40
0 F1 (Hz)
-40 FIGURE 8-50 The proton decoupled,

(C)
I I I +40 two-dimensional J-resolved 13C spectra
I I
I I of a mixture of methyl deuterated
I I toluenes (2 H has an / = 1). J coupling is
I I g I +20
I I
shown on the F1-axis and the
I I corresponding 13C Shifts on the F2- axis.
The single peak for CH 3 , triplet for
* I -0 F1(Hz)
-CH 2D, pentet for -CHD 2, and septet
I I for CD3 are seen on the F1 axis for slices
I I I I -20 through F2 corresponding to the
I I analogous 13C chemical shift. The
I
I
I
I I one-dimensional 13C spectrum is shown
I I I
-- 40 in (A). Dashed lines connect the
I I
chemical shifts of the two-dimensional
30 60 90 and one-dimensional spectrum.
F2(z)
(B)
CH3
from phosphorus-proton couplings. The reason for this is that if the proton-proton
coupling is modulated during ti, the phosphorus-proton couplings remain and
are present on the v2-axis. An example is illustrated in Fig. 8-51(A) for the complex
Cp-Ru-(CH=CH 2)(P(C6 H) 3 ) 2. The spectrum of the proton on the vinyl
group of the carbon bound to ruthenium is shown in Fig. 8-51(A). This should
consist of overlapping triplets from two equivalent phosphorus atoms split into
a doublet of doublets from non-equivalent CH 2 protons of the vinyl group. The
2-D, J-resolved, 'H nmr spectrum shows the phosphorus splitting on the F,(6)
axis, as seen by the triplet in the spectrum in Fig. 8-51(C). The F 2 axis (J)contains
the information about the proton-proton coupling. The doublet of doublets is
indicated. Using the resolved coupling constants, a stick diagram can be con-
structed to show how the overlap of peaks leads to the spectra in Fig. 8-51(A).
+15
-0 O(Hz)
FIGURE 8-51 (A) The proton spectrum

of the vinyl CH proton of CpRu(CH=CH 2)
(P<3)2. (B) The J-resolved,
two-dimensional spectrum. (C) The
proton spectra with 31P splittings. Cp Ru ,/ PPh 3
PPh 3
3
H 2 H
H2
(C)
NMR in Solids 8-20 DIRECT DIPOLAR COUPLING

and Liquid We shall begin this discussion by considering the direct dipole-dipole interaction
Crystals between two nuclei. This is done by consideringfixed orientationsof two hydrogen
atoms in a molecule relative to an external field, as shown in Fig. 8-52. The
dashed line is the internuclear axis connecting the two hydrogen atoms, the
boldface arrow indicates the orientation of the nuclear moment on b relative to
the field, and the curved arrow represents the lines of flux arising from this nuclear
moment. We see that for this fixed orientation of the molecule, we would obtain
two different peaks in the nmr of the Ha resonance as a consequence of the two
different fields from b. This is a through-space effect which, in contrast to the
coupling mechanisms discussed earlier, does not involve the electron density in
the molecule. The peak separation is indicated by a coupling constant Bdi, (where
dir stands for direct). B can be very large and, for example, is about 120,000 Hz
for two protons separated by 1 A when the H-H internuclear axis is aligned
with the field.
(A) (B)
FIGURE 8-52 The direct dipolar interaction of two
hydrogens in a molecule.
8-21 NMR Studies of Solids 341
The magnitude of B will vary with the position of the internuclear axis in
relation to the field. The mathematical expression for the magnitude of the direct
dipolar interaction between two nuclei p and q, Bpqir, is derived from the
expression for the magnetic field due to a point magnetic dipole (i.e., the other
nucleus) and the expression for the potential energy of a point dipole (the proton
of interest) in a magnetic field. The result is:
. hi 1 [3 COS2
2 gp _ 1
1 r, (8-26)
Bd'(Hz) =
pq47
h4
i~
q 23
2 L rpq 3
34
where pq is the angle that the pq-internuclear axis makes with the external field
direction. For a complicated molecule there is one Bdir for every pair of magnetic
nuclei in the sample.
In a normal solvent, any given internuclear axis is randomly oriented with
respect to the external field and rapidly changing its position because the molecule
is rotating. The direct dipole-dipole interaction is averaged to zero. Since the
indirect coupling constants proceed through the electron density of the molecule,
they are not averaged to zero in a solvent. However, the direct dipole-dipole
interaction is through space: it is given by equation (8-26) and is averaged out.
At very high fields (,> 14 T), the effects of imperfectly averaged diamagnetic
anisotropy are seen, e.g., broadening of aromatic proton peaks.
8-21 NMR STUDIES OF SOLIDS
In single crystals, the H-H internuclear axes in the molecules have a fixed
orientation relative to the applied field and B dir is not averaged out. According
to equation (8-26), if one were to study the angular variation of B'ir by
investigating different orientations of a single crystal, it would be possible to
solve the resulting simultaneous equations for 1/r3 and to find the magnitude
and direction of the H-H internuclear axes. Unfortunately, only rarely are the
protons in the crystal few enough in number and far enough apart to permit
resolution of the spectral lines and determination of B dir. Generally, the spectra
of solids consist of very broad, poorly resolved bands because of the direct dipolar
interaction between protons in the molecule and between those from the nearby
molecules. However, structural information can be obtained from the broadened
resonance line of single crystals or powders by the so-called method of second
2
moments( 66). The second moment is the mean square width (AH) measured from
the center of the resonance line. The center of the resonance line is the average
magnetic field, as seen in equation (8-27).
Hay = { Hf(H) dH (8-27)
where f(H) represents the normalized line shape. The second moment is then
given by equation (8-28),
(AH) 2 = (H - Hav) 2 f(H) dH (8-28)

which can be evaluated graphically from the observed line shape. Most observed
peaks have a Gaussian shape and are described by
1
f(H) = __ e- H )/2(A H) (8-29)
AHI2r
The second moment is then determined from the integration of the analytical
expression obtained from equations (8-28) and (8-29). The second moment for
a single crystal or powder is related to r and 0, and gives structural
information. The reader is referred to references 66 through 68, for more details.
This technique has been used to show that the infusible white precipitate from
the reaction of NH 3 with HgC 2 is NH 2HgCI and not NHg 2ClNH 4 Cl or
XHgO(1 - X) HgCl2 -2NH 3.1 691 These studies can be carried out on more
complex molecules in conjunction with deuteration studies. The magnetic moment
of the deuteron is very small, and dipolar interactions that involve deuterium
can usually be neglected.
The temperature dependence of the second moment has also been employed
to provide information on molecular motion in solids. It has been shown that
benzene'is fixed in the solid below 90 K, but rotates rapidly about the sixfold
axis between 120 and 280 K. The second moment changes gradually from 9.7
gauss 2 below 90 K to 1.6 gauss 2 at 120 K, and remains at this value to 280 K.
Rotation of cyclohexane about the S, axis has also been demonstrated. This is
an extremely sensitive technique, for the rate of rotation required to narrow the
resonance is not much higher than the proton resonance frequency in a field of
a few gauss, i.e., 10' Hz. Second-moment studies demonstrated( 70 ) the rotation
of the benzene and cyclopentadiene rings in dibenzene-chromium and ferrocene
as well as(71) rotation of benzene in the benzene-silver perchlorate complex. A
study( 72 ) of [Me 3SiNSiMe2] 2 demonstrated that at 77 K the methyl groups rotate
about the Si-C bond, and this is the only motion in the molecule. At room
temperature, methyl groups rotate about the C-Si bond, (CH 3 )3 Si groups rotate
about the Si-N bond, and the whole molecule rotates about a molecular axis.
Thermodynamic data can be obtained for the various motional processes from
the temperature dependence of the spectrum.
8-22 NMR STUDIES IN LIQUID CRYSTAL SOLVENTS
Certain materials, e.g.,
0
CH 3CH 20 N=N OC(CH 2 )5 CH3
and
CH 3 (CH 2 )6 N=N O(CH 2)6 CH 3
melt to produce a turbid fluid in which certain domains exist where there is
considerable ordering of the molecules. The resulting fluid has some properties
of both liquids and crystals. The liquid is strongly anisotropic in many of its
properties,73 1 but further heating produces an isotropic liquid. This general class
of materials is referred to as thermotropicliquid crystals (i.e., produced by melting
a solid). Three subclassifications are illustrated in Fig. 8-53 by schematically
Nematic Smectic
FIGURE 8-53 Types of ordered domains in a liquid crystal.
indicating the type of ordering in the domains. The representations in Fig. 8-53
are of small domains in the whole liquid sample. These domains tend to become
aligned in the presence of a magnetic field.
The nmr spectra of the molecules that constitute the liquid crystal are very
broad, nondescript resonances that sometimes are so broad as not to be observed.
The broadening results because the viscous solvent molecules contain many
hydrogens and consequently there are a large number of direct H-H dipole-dipole
interactions that are not averaged out in the partially ordered solvent.
If some benzene is dissolved in the liquid crystal as solvent, the spectrum
illustrated in Fig. 8-54(A) results.(7" This spectrum of benzene consists of a large
number of sharp lines spread out over approximately 2400 Hz. Some of the broad
resonances of the solvent are discernible under the sharp benzene peaks. The
benzene spectrum is thus somewhere between that obtained in a non-viscous
solvent (a single sharp line) and that of a solid (a broad line spread over a wide
-1000 Hz-
FIGURE 8-54 (A) Liquid crystal nmr spectrum of

benzene; (B) simulated spectrum.
field). In an isotropic solvent, the various axes in a solute molecule assume all
possible orientations with respect to the magnetic field with equal probability,
and B is averaged to zero. As mentioned before, the indirect couplings proceed
via the electron density in the molecule and are not averaged to zero. When all
orientations of an axis with respect to the field are not equally probable, the
contribution to the field at a proton from each surrounding nucleus depends on
the orientation of the nuclear moment in the neighbor, and the position of that
neighbor nucleus relative to the one being observed. In a solid where the nuclei
are fixed and where so many protons affect the field of any one kind of proton,
a continuous very broad single peak results. We now must explain why we see
such sharp peaks for the solute in the liquid crystal spectrum and why there are
so many peaks.
In a liquid crystal, certain orientations of the solute are more favored than
others (long axes line up with the solvent long axis) because the magnetic field
tends to align the solvent and solute molecules in the field direction. Solute
molecules diffuse freely and tumble freely enough so that there is no contribution
to Bdir from the solvent or other solute molecules. All the couplings are
intramolecular. Since the molecule is tumbling rapidly, sharp lines are observed
in the nmr. We see many lines in the spectrum because Bpqdir [equation (8-26)]
makes an observed contribution to the resonance line positions. The anisotropic
motion does not average this quantity to zero. Each Oq of equation (8-26) now
becomes some average value for the net orientation of each of the respective H-H
axes in this rapidly and anisotropically tumbling molecule.
In benzene, for example, we have six protons and need a matrix of all possible
combinations of six spins (i.e., a basis set 222222, oc22fl, etc.) to describe this
system. Thus, we have peaks corresponding to molecules with all these different
permutations of spins with B and J values for all pairs of hydrogens. Often,
second-order spectra result, further complicating the appearance of the spectrum.
The energies are described by the spin Hamiltonian matrix for this system, which
is similar to that discussed in treating second-order spectra except that, for every
J on the diagonal in the solution problem, we now have a J + 2B; and for every
J previously on the off diagonal we now have a J - B. This problem is solved,
vide infra. by finding J and B values [Jortho(o)Jmeta(m), Jpara(p), B0 , B., and B,]
that will reproduce the experimental spectrum. The calculated spectrum is shown
in Fig. 8-54(B). For benzene, the resulting values are B. = -639.5 Hz,
B.= -123.1 Hz, B = - 79.93 Hz, J. = 6.0 Hz, J.= 2.0 Hz, and J, = 1.0 Hz.
The dipolar couplings are related to the nuclear coordinates by
3 2
(Bpq> -8 7q [ ,,(3 'cos
p, _ 1)
+ (S - S,,) (cos 2 pq, - cos 2 opq,) + 4S,,cos 0 PqXcos 0

pqy
4 0
+ Sx, cos pqx COSOpqz + 4S Cos Pqy COs pqzl (8-30)
where rq is the internuclear separation of the coupled nuclei, pqz is the angle rpq
makes with the z-coordinate, etc. for 0 Pqx and pqy. The axes are fixed within the
molecule and S is an averaged-order matrix that relates to the extent of orientation
of the molecule by the liquid crystal. If the axis is completely aligned with the
field, SPq is 1. If the axis is aligned perpendicular to the field, then Spq = (/2.For
a random orientation, Spq = 0.
If we knew SP, we could calculate rpq from the experimental Bpqdir and get
distances and geometries of molecules in solution. In a rigid molecule, the values
of Spq for the various axes must be interrelated. This interrelationship between
the axes and the orientation of the molecule in three-dimensions can be described
with the tensor 8. The S tensor is 3 x 3, symmetric and traceless, so that only
five of the nine elements are independent. As discussed in Chapter 2, we can
define a molecular coordinate system that diagonalizes the tensor. This usually
corresponds to symmetry axes if the molecule has symmetry; then there will be
only two independent elements. (All off-diagonal elements are zero and the
diagonal ones are traceless, i.e., their sum equals zero.) For a molecule with a
threefold or higher axis, S., = S,, = -('/ 2 )S2 and there is only one independent
tensor element. Note that the tensor is independent of r. With a threefold or
high-order axis, equation (8-30) becomes
B Bdir' -h S,(3 cos 2 O, 1)

(8-31)
8x 4 rq32
B0
We can illustrate the method by considering paradichlorobenzene, Fig. 8-55,
as an example. Analysis of the spectra yields three dipolar coupling constants,
B0, B., and B, shown in Fig. 8-55. There are two unknown tensor elements S,
and (S., - S,,) and two distances corresponding to the two sides of the triangle
made by B,, B., and H-H in Fig. 8-55. We are thus confronted with three
equations and four unknowns. Therefore, it is only possible to get the ratios of
all the distances; or, if we can assume one, the others can be calculated. It is
often the case in these systems that there is one more unknown than there are
knowns. However, even if a particular system has more coupling constants than
unknowns, we cannot obtain a unique solution. This can be seen by referring to FIGURE 8-55 The Bp
equation (8-31). For a given S and r that satisfy equation (8-31) we could multiply interactions in p-Cl 2 CH 4.
r by a factor and S by the cube of that factor and get the same B. This corresponds
to a uniform expansion of the structure without any change in the ratios of the
distances. In spite of this severe limitation, a very considerable amount of
information can still be obtained from nmr studies in liquid crystals.
In molecules with a threefold axis or higher, equation (8-31) applies. These
systems require only one S element to relate the Bpqvalues to structure. The ratios
of the dipolar coupling constants are independent of S and give us the shape of
the molecule. The ratio of the dipolar coupling for all proton pair Bpq values can
be calculated for a regular hexagon by taking the corresponding ratios of equation
(8-31) to yield I to 0.192 to 0.125. Experimental ratios of I to 0.192 to 0.125 are
found for benzene, showing that it is a regular hexagon in solution. CpNiNO
has two Bq values whose ratio for a regular pentagon is 4.236. A value of 4.211
is found indicating a regular polygon. On the other hand, CpMn(CO) 3 produces
an experimental ratio of 4.11 indicating that distortion exists.( 75'7 61 In a study 77 )
of the liquid crystal nmr spectrum of trans-[Cl2 Pt(C2 H4 )C5 HN], the influences
that various dynamic processes in solution have on the resulting spectra are
discussed. The ordering factor S is very low and severe overlapping of the
resonances results. A method was developed in which the spectrum intensities
are fitted in the spectral analysis. The ratio of the r.,g/ri, protons in the
coordinated ethylene is consistent with a structure in which the C-H bonds are
bent back away from the metal. Except for the dynamic process occurring in
solution (the ethylene rotates rapidly about the bond to platinum), the solution
structure of the ethylene fragment is similar to that in the solid.
When the chemical shift differences of the protons involved are large and
first-order spectra are obtained, the spectrum is readily interpreted by inspection,
and information about the symmetry of the molecule can be obtained from the
spectrum in a straightforward manner. Consider the results reported(17 8) for
H3Ru 3(CO)CH3 , which led to the proposed structure shown in Fig. 8-56 along
with the spectrum observed for this molecule in a liquid crystal. The methyl
00 0
CC C
\l/
Ru
H H
CHa
C
OC II-" 11 CO
OC-Ru Ru-CO
FIGURE 8-56 (A) Structure OC H ,CO
of HaRu,(CO)CH 3. (B) (A)
Experimental liquid crystal
spectrum (broad absorption
is from the liquid crystal). (C)
Simulated liquid crystal
spectrum. (Reprinted with
permission from A. D.
Buckingham, et al., J. Amer.
Chem. Soc., 95, 2732 (1973).
Copyright by the American
Chemical Society.]
200 Hz
11
111,
11
group is rapidly rotating about the C-C bond axis. The quartets arise from the
three equivalent methyl protons splitting the three bridging protons and vice
versa. The three methyl protons are equivalent with respect to the bridging
hydrides because of the rapid rotation. In a fixed staggered configuration (i.e.,
CH 3 versus metal hydrides), the hydride protons would be split by a set of two
equivalent protons and one non-equivalent methyl proton, producing six peaks
in each triplet component, i.e., 36 peaks total. One of the two groups of triplets
arises from direct dipolar coupling of the methyl protons with each other, and
the second (more closely spaced) triplet arises from the dipolar coupling of the
three bridged hydrides with each other. If we consider one of the protons of the
8-23 High Resolution NMR of Solids 347
methyl group, it can be split by the other two because (+ +),(+ -), (- +), and
(- -) combinations of nuclear moments in different molecules cause this proton
to experience different fields, since the molecule is not undergoing isotropic
rotation and the direct dipolar coupling is not averaged to zero. The larger
coupling constants arise when the protons causing the splitting are closer to each
other.
Deuterium nmr in liquid crystals 8 -7 9 ) leads to a determination of the B
quadrupole coupling constant qz. (Chapter 14). The interaction of a proton with
a deuteron in a partially aligned molecule will give rise to a four-line pattern.
The large doublet arises from the quadrupole interaction and the smaller doublet
from the dipolar coupling. The value of q,,, is related to the quadrupole coupling - Sp, qN
constant by FIGURE 8-57 Splitting of
the deuterium resonance by
qpzz = Spqqzz (8-32) a proton.
8-23 HIGH RESOLUTION NMR OF SOLIDS 80 )
As described in Section 8-21, dipolar coupling leads to very broad nmr spectra
of solids. Also contributing to the broadness is chemical shift and scalar spin-spin
coupling (J) anisotropy with both quantities being represented by tensors. The
entire range of chemical shifts and couplings spanned by all orientations of
crystallites is manifested in the spectra leading to widths of 50-200 ppm in "C
spectra. The first attempts to reduce dipolar coupling in solids involved rapid
spinning of the sample about an axis, inclined at the so-called magic angle of
54 44'. All of the moment vectors are pointing in all directions in the static
sample. One such moment is illustrated by the vector pq relative to H, in Fig.
8-58. If we spin rapidly, we average all of these leading to a net moment that
points along the spinning axis. This can be seen in Fig. 8-58 where spinning
along the axis at 54 44' sweeps out two circles for opposite ends of pq. The net
of all the orientations along the circles is the vector p'q' lying on this spinning FIGURE 8-58 Averaging of
axis. At this magic angle for the axis, 3 cos 2 0- 1 equals zero. One must spin the moment pq to p'q' by
the sample rapidly compared to the magnitude of the interaction to average the spinning about the axis at
moments. Dipolar coupling of two protons typically is 5 x 10' Hz. Centrifugal 54 44' to the applied He field.
force will cause rotors to disintegrate at this speed. Conventional spinning rates
are fast enough to remove chemical shift and J coupling anisotropy. Large
chemical shift anisotropy produces severe spinning sidebands.
This approach, which will be called magic angle sample spinning, MASS,
can remove dipolar coupling in magnetically dilute samples, e.g., "C. Note in
3
equation (8-26) that the direct coupling is a function of 1/r . Also note (equation
(8-26)) that a low y value for the nucleus of interest also tends to reduce dipolar
coupling. Thus, "C and 2 Si give sharp nmr spectra from. MASS.
Quadrupolar nuclei also lead to line broadening through the electric
quadrupole interaction. For I = n + '/2 nuclei, the center transition (-'/2 -+
'/2) is not affected and a sharp line results. The other transitions are generally
broadened beyond detection. For nuclei with integer values of I, there is no
- '/2 - '/2 transition and very broad spectra result. Even with I = n/2 systems,
a second-order quadrupole effect broadens the central transition. This effect
becomes less important at high fields.
Low values of y and chemical dilution permits detection of 100% "N and
"P in solids with broadband proton decoupling. Long T's can be shortened by
cross polarizing,(81 ) CP, dilute, and abundant spins. In the CP technique, the
(A) proton spin polarization is transferred(s2 ) to the less abundant nuclei, e.g., "c
by applying a proton rf field (Hi,,) and then a 3 C field (H,) such that
yHHiH -- ycHc. Efficient transfer of energy occurs between the two spin manifolds
generating 3 C magnetization along Hc and increasing the intensity of the
transition. Cross polarization depends on the static component of the dipolar
interaction and has the greatest effect on neighboring nuclei. It is most efficient
for immobile species and ineffective for mobile species. These concepts have been
employed(8 3,84 ) to study the surface structure and reactivity of silica gel and
zeolites. Complementary information about the bulk structure can be obtained
using a pulse-delay-observe sequence.
(C)
Combining chemical dilution with MASS leads to a further sharpening of
FIGURE 8-59 C nmr of the resonances as shown in Fig. 8-59. A pulse sequence referred to as
polysulfone WAHUHA ss> modulates the spin and has reduced line widths from 104 to - 102
(-O-C(CH 3) -0-O-O-
2 Hz. These experiments have the potential for determining chemical shift and
S02-0-O-) (A) double dipolar anisotropes. Carbon-hydrogen dipolar couplings and with them C-H
resonance; (B) double
resonance plus MASS; (C)
distances have been obtained on solids by a 2-D pulse sequence.("6 ) The evolution
solution spectrum. period corresponds to variable times, r, for the application of a WAHUHA pulse.
The 2-D plot contains the chemical shift as one axis and the dipolar splittings
as the other.
REFERENCES 1. F. L. Slejko, R. S. Drago and D. G. Brown, J. Amer. Chem. Soc., 94, 9210 (1972).
CITED 2. S. S. Zumdahl and R. S. Drago, J. Amer. Chem. Soc., 89, 4319 (1967).
3. S. Alexander, J. Chem. Phys., 37, 967 (1962); Rev. Mol. Physics, 24, 74 (1957).
4. C. S. Johnson, Adv. Mag. Res., 1, 33 (1965).
5. G. Fano, Rev. Mod. Physics, 24, 74 (1957).
6. S. S. Zumdahl and R. S. Drago, Inorg. Chem., 7, 2162 (1968).
7. (a) W. D. Perry, R. S. Drago and N. K. Kildahl, J. Coord. Chem., 3, 203 (1974).
(b) N. K. Kildahl and R. S. Drago, J. Amer. Chem. Soc., 95, 6245 (1973).
8. H. S. Gutowsky and C. H. Holm, J. Chem. Phys., 25, 1228 (1956).
9. M. Rogers and J. C. Woodbrey, J. Phys. Chem., 66, 540 (1962).
10. C. Looney, W. D. Phillips and E. L. Reilly, J. Amer. Chem. Soc., 79, 6136 (1957).
11. G. Claeson, G. Androes and M. Calvin, J. Amer. Chem. Soc., 83, 4357 (1961).
12. E. Grunwald, A. Loewenstein and S. Meiboom, J. Chem. Phys., 27, 630 (1957).
13. S. Meiboom et al., J. Chem. Phys., 29, 969 (1958) and references therein.
14. E. L. Muetterties and D. W. Phillips, J. Amer. Chem. Soc., 81, 1084 (1959).
15. A. Patterson, Jr. and R. Ettinger, Z. Elektrochem., 64, 98 (1960).
16. W. D. Phillips et al., J. Amer. Chem. Soc., 81, 4496 (1959).
17. W. N. Lipscomb, Adv. Inorg. Chem. Radiochem., 1, 132 (1959).
18. B. M. Graybill, J. K. Ruff and M. F. Hawthorne, J. Amer. Chem. Soc., 83, 2669 (1961).
19. M. J. Bennett, Jr., et al., J. Amer. Chem. Soc., 88, 4371 (1966); F. A. Cotton, Accts.
Chem. Res., 1, 257 (1968).
20. (a) K. Vrieze and P. W. N. M. van Leeueven, Prog. Inorg. Chem., 14, 1 (1971).
(b) "Dynamic Nuclear Magnetic Resonance Spectroscopy," ed. L. M. Jackman and
F. A. Cotton, Academic Press, New York, 1975.
21. L. W. Reeves and K. N. Shaw, Can. J. Chem., 48, 3641 (1970).
22. H. S. Gutowsky and C. J. Hoffman, J. Chem. Phys., 19, 1259 (1951).
23. J. I. Musher, J. Amer. Chem. Soc., 94, 5662 (1972) and references therein.
24. (a) W. G. Klemperer, J. Amer. Chem. Soc., 95, 380 (1973) and references therein.
(b) J. P. Jesson and P. Meakin, J. Amer. Chem. Soc., 96, 5760 (1974).
25. G. M. Whitesides and H. L. Mitchell, J. Amer. Chem. Soc., 91, 5384 (1969).
26. J. R. Shapley and J. A. Osborn, Accts. Chem. Res., 6, 305 (1973).
27. J. P. Jesson, in "Transition Metal Hydrides," ed. E. L. Muetterties, M. Dekker, New
York, 1971.
28. S. S. Eaton, J. R. Hutchison, R. H. Holm and E. L. Muetterties, J. Amer. Chem. Soc.,
94, 6411 (1972).
29. D. J. Duffy and L. H. Pignolet, Inorg. Chem., 11, 2843 (1972).
30. J. W. Faller, in "Determination of Organic Structures by Physical Methods," Vol. 5,
eds. F. Nachod and J. Zuckerman, Academic Press, New York, 1973, and references
therein.
31. (a) P. D. Kennewell, J. Chem. Educ., 47, 278 (1970).
(b) See, for example, A. Carrington and A. D. McLachlan, "Introduction to Magnetic
Resonance," Chapter 13, Harper & Row, New York, 1967.
(c) J. H. Noggle and R. E. Schermer, "The Nuclear Overhauser Effect: Chemical
Applications," Academic Press, New York 1971.
(d) D. Neuhaus, P. Davis, and M. Williamson, "The Nuclear Overhauser Effect in
Stereochemical and Conformational Analysis," VCH Publishers, New York, 1989.
32. G. C. Martin and J. M. Boncella, Organomet., 8, 2969, (1989).
33. (a) G. L. Closs et al., J. Am. Chem. Soc., 92, 2183, 2185, 2186 (1970).
(b) S. H. Pine, J. Chem. Educ., 49, 664 (1972).
34. T. C. Farrar and E. D. Becker, "Introduction to Pulse and Fourier Transform NMR
Methods," Academic Press, New York, 1971.
35. C. Brevard and P. Granger, "Handbook of High Resolution Multinuclear NMR,"
Wiley, New York, 1981.
36. E. D. Becker, J. A. Feretti, M. Gambhir, Anal. Chem. 51, 1413 (1979).
37. K. L. Kuhlmann, D. M. Grant and R. K. Harris, J. Chem. Phys., 52, 3439 (1970).
38. T. D. Alger, S. W. Collins and D. M. Grant, J. Chem. Phys., 54, 2820 (1971).
39. A. Allerhand and D. Doddrell, J. Am. Chem. Soc., 93, 2777 (1971).
40. R. A. Newmark and J. R. Hill, J. Am. Chem. Soc., 95, 4435 (1973).
41. (a) 0. A. Gansow, A. R. Burke and G. N. LaMar, Chem. Comm., 456 (1972).
(b) 0. A. Gansow et. al., J. Amer. Chem. Soc., 94, 2550 (1972).
(c) D. F. S. Natusch, J. Amer. Chem. Soc., 93, 2566 (1971).
42. D. Doddrell and A. Allerhand, Proc. Nat]. Acad. Sci. (USA), 68, 1083 (1971).
43. 0. A. Gansow and W. Schittenhelm, J. Amer. Chem. Soc., 93, 4294 (1971).
44. N. A. Matwiyoff and D. G. Ott, Science, 181, 1125 (1973).
45. P. J. Vergamini, N. A. Matwiyoff, R. C. Wohl and T. Bradley, Biochem. Biophys. Res.
Comm., 55, 453 (1973).
46. N. A. Matwiyoff et al., J. Amer. Chem. Soc., 95, 4429 (1973).
47. R. E. London, C. T. Gregg and N. A. Matwiyoff, Science, 188, 266 (1975).
48. R. F. London, V. H. Kollman and N. A. Matwiyoff, J. Amer. Chem. Soc., 97, 3565
(1975).
49. G. A. Olah et al., J. Amer. Chem., Soc., 93, 4219 (1971).
50. R. G. Parker and J. D. Roberts, J. Amer. Chem. Soc., 92, 743 (1970).
51. G. M. Bodner et al., Chem. Comm., 1970, 1530.
52. G. M. Bodner and L. J. Todd, Inorg. Chem., 13, 360 (1974); ibid., 13, 1335 (1974).
53. L. J. Todd and J. R. Wilkinson, J. Org. Met. Chem., 77, 1 (1974).
54. B. E. Mann, Adv. Org. Met. Chem., 12, 135 (1974).
55. J. Evans and J. R. Norton, Inorg. Chem., 13, 3043 (1974).
56. R. Ditchfield and P. D. Ellis, in "Topics in Carbon-13 NMR Spectroscopy," Vol. 1,
ed. G. C. Levy, Wiley-Interscience, New York, 1974.
57. R. K. Harris and B. E. Mann, "NMR and the Periodic Table," Academic Press, New
York, 1978.
58. J. Mason, Chem. Brit., 654 (1983).

59. (a) I. C. P. Smith, G. W. Stockton, A. P. Tulloch, C. F. Polnaszek and K. G. Johnson,
J. Colloid Interface Sci., 58, 439 (1977).
(b) A. Seelig and J. Seelig, Biochemistry, 13, 4839 (1974).
(c) H. H. Mantsch, H. Sait6 and 1. C. P. Smith, in "Progress in Nuclear Magnetic
Resonance Spectroscopy," ed. J. W. Emsley, J. Feeney and L. H. Sutcliffe, Pergamon
Press, London, 1977.
60. H. Schmidbaur, 0. Gasser, T. E. Fraser and E. A. V. Ebsworth, J. Chem. Soc., Chem.
Comm. 34 (1971).
61. A. Derome, "Modern NMR Techniques for Chemistry Research," Pergamon Press,
London, 1987.
62. D. M. Dodrell, D. T. Pegg, M. R. Bendall, J. Mag. Reson., 48, 323 (1982).
63. T. L. Venable, W. C. Hulton and R. N. Grimes, J. Am. Chem. Soc., 106, 29 (1984).
64. (a) C. Brevard, R. Schimpe, G. Tourne and C. M. Tourne, J. Am. Chem. Soc., 105,
7059 (1983).
(b) R. G. Fuke, B. Rapho, R. J. Sexton and P. J. Vomaille, J. Am. Chem. Soc., 108,
2947 (1986).
65. See, for example, N. W. Alcoch, J. M. Brown, A. E. Derome and A. R. Lucy, J. Chem.
Soc., Chem. Comm., 757, (1985).
66. C. P. Slichter, "Principles of Magnetic Resonance," Chapter 3, Harper & Row, New
York, 1963.
67. A. Abragam, "The Principles of Nuclear Magnetism," Chapter 4, Clarendon Press,
Oxford, 1961.
68. E. R. Andrew, "Nuclear Magnetic Resonance," Chapter 6, Cambridge University
Press, New York, 1955; E. R. Andrew and P. S. Allen, J. Chem. Phys., 63(1), 85-91
(1966).
69. C. M. Deely and R. E. Richards, J. Chem. Soc., 3697 (1954).
70. L. N. Mulay, E. G. Rochow and E. 0. Fischer, J. Inorg. Nucl. Chem. 4, 231 (1957).
71. D. F. R. Gilson and C. A. McDowell, J. Chem. Phys., 40, 2413 (1964) and references
therein.
72. H. Levy and W. E. Grizzle, J. Chem. Phys., 45, 1954 (1966).
73. P. Diehl, et al., "NMR Basic Principles and Progress," Vol. 1, Springer-Verlag, New
York, 1969.
74. L. C. Snyder and E. W. Anderson, J. Am. Chem. Soc., 86, 5023 (1964).
75. J. W. Emsley and C. S. Lunon, Mol. Phys., 28, 1373 (1974).
76. C. S. Yannoni, G. P. Ceasar and B. P. Dailey, J. Am. Chem. Soc., 89, 2833 (1967).
77. D. R. McMillin and R. S. Drago, Inorg. Chem. 13, 546 (1974).
78. P. Diehl and A. J. Tracey, Molec. Phys., 30, 1917 (1975).
79. J. W. Emsley et al., J. Chem. Soc., Faraday II, 1365 (1976).
80. R. K. Harris, "Nuclear Magnetic Resonance Spectroscopy," Pitman, London, 1983.
81. A. Pines, M. Gibby and J. S. Waugh, J. Chem. Phys. 59, 569 (1973) and references
therein.
82. S. R. Hartman and E. C. Halsk, Phys. Rev, 128, 2042 (1962).
83. D. W. Sindorf and G. E. Maciel, J. Am. Chem. Soc. 105, 1487, 3767 (1983).
84. C. A. Fyfe, G. C. Gobbi, J. Klinowski, J. M. Thomas and S. Ramdas, Nature 296,
530 (1982); J. Phys. Chem. 86, 1247 (1982).
85. J. S. Waugh, L. M. Huber and U. Haeberlen, Phys. Rev. Lett., 20, 180 (1968).
86. J. S. Waugh, Proc. Nat]. Acad. Sci., USA, 73, 1394 (1976); R. K. Hister, J. L. Ackerman,
B. Neff and J. S. Waugh, Phys. Rev. Lett. 36, 1081 (1976).
J. W. Emsley et al., "High Resolution Nuclear Magnetic Resonance Spectroscopy," Vols, ADDITIONAL
I and 2, Pergamon Press, New York, 1966.
T. C. Farrar and E. D. Becker, "Introduction to Pulse and Fourier Transform NMR REFERENCES
Methods," Academic Press, New York, 1971.
C. P. Slichter, "Principles of Magnetic Resonance," Harper & Row, New York, 1963.
A. Abragam, "The Principles of Nuclear Magnetism," Clarendon Press, Oxford, 1961.
G. C. Levy and G. L. Nelson, "Carbon-13 Nuclear Magnetic Resonance for Organic
Chemists," Wiley-Interscience, New York, 1972.
J. B. Stothers, "Carbon-13 NMR Spectroscopy," Academic Press, New York, 1972.
I. Bertini, H. Mollnari and N. Niccolai, "NMR and Biomolecular Structure," VCH
Publishers, New York, 1991.
"Nuclear Magnetic Resonance Shift Reagents," ed. R. E. Sievers, Academic Press, New
York, 1973.
C. P. Poole, Jr. and H. A. Farach, "The Theory of Magnetic Resonance," Wiley-
Interscience, New York, 1972.
E. Breitmaier and W. Voelter, "Carbon-I 3 NMR Spectroscopy: High-Resolution Methods
and Applications in Organic Chemistry and Biochemistry," VCH Publishing, New
York, 1987.
A. Derome, "Modern NMR Techniques for Chemistry Research," Pergamon Press,
London, 1987.
G. E. Martin and A. S. Zektzer, "Two-Dimensional NMR Methods for Establishing
Molecular Connectivity: A Chemist's Guide to Experiment Selection, Performance
and Interpretation," VCH Publishing, New York, 1989.
"Two-Dimensional NMR Spectroscopy: Applications for Chemists and Biochemists," eds.
W. R. Croasmun and R. M. K. Carlson, VCH Publishing, New York, 1987.
A. E. Tonelli, "NMR Spectroscopy and Polymer Microstructure: The Conformational
Connection," VCH Publishing, New York, 1989.
W. S. Brey, ed. "Pulse Methods in 1-D and 2-D Liquid Phase NMR," Academic Press,
New York, 1988.
R. Freeman, "A Handbook of Nuclear Magnetic Resonance," Wiley, New York, 1987.
R. R. Ernst, G. Bodenhausen and A. Wokaun, "Principles of Nuclear Magnetic Resonance
in One and Two Dimensions," Oxford University Press, Oxford, 1987.
A series of articles by R. W. King and K. R. Williams provides a clear explanation of FT
methods:
"The Fourier Transform in Chemistry. Nuclear Magnetic Resonance: Introduction," J.
Chem. Educ., 66, A213-A219 (1989).
"The Fourier Transform in Chemistry. Part 2. Nuclear Magnetic Resonance: The Single
Pulse Experiment," J. Chem. Educ., 66, A243-A248 (1989).
"The Fourier Transform in Chemistry. Part 3. Nuclear Magnetic Resonance: Multiple
Pulse Experiments," J. Chem. Educ., 67, A93-A99 (1990).
"The Fourier Transform in Chemistry. Nuclear Magnetic Resonance: A Glossary of N MR
Terms," J. Chem. Educ., 67, A100-A104 (1990).
"The Fourier Transform in Chemistry. Part 4. Nuclear Magnetic Resonance: Two-
Dimensional Methods," J. Chem. Educ., 67, A125-A137 (1990).
Series
"Annual Reports on NMR Spectroscopy," Academic Press, New York, 1968-
"Advances in Magnetic Resonance," ed. by J. S. Waugh, Academic Press, New York, 1965-
"Advances in Magnetic Resonance, Supplement 1: High Resolution NMR in Solids," U.
Haeberlen, Academic Press, New York, 1976.
"NMR Abstracts and Index," Preston Technical Abstracts Co., Evanston, Ill., 1968.
"Nuclear Magnetic Resonance," The Chemical Society (Specialist Periodical Report),

London, 1972-
"Topics in Carbon-13 NMR Spectroscopy," ed. G. C. Levy, Wiley-Interscience, New York,
1974-.
EXERCISES 1.Consider the molecule
M M
X -X " ,X
where M and X have I 2 Sketch the nmr spectra for the following conditions,
assuming A > J in all cases (when necessary to assume orders of magnitude for various
coupling constants, state your assumption). Ignore coupling if the atoms are not
directly bonded.
a. nmr spectrum of X, no exchange.
b. nmr spectrum of M, no exchange.
c. nmr spectrum of X with rapid intermolecular exchange of all X groups.
d. nmr spectrum of M with rapid [T' < /(vA - VB) intermolecular exchange of all X
groups.
e. nmr spectrum of X with rapid intramolecular exchange.
f. nmr spectrum of M with rapid intramolecular exchange.
2. In the absence of any exchange, two peaks A-H and B--H are separated by 250 Hz.
At room temperature, exchange occurs and the peaks are separated by 25 Hz. The
spin-lattice relaxation of A-H and B-H is long, and there are equal concentrations
(0.2 M) of each. Calculate the lifetime of a proton on A, and from this the rate constant
for the exchange (specify units).
3. In a given compound MF 4 (for M, I = 2) the JM-F value is 150 Hz. In the absence
of chemical exchange, the F- and M-F signals are separated by 400 Hz. At room
temperature the F- and MF 4 exchange at a rate such that the fine structure just
disappears. Assuming equal concentrations of M F and F- species and no stable
intermediates, calculate T'for F~. What will be the separation of the M-F and F-
peaks under these conditions?
4. The following are 3P nmr spectra, I, = IF 2. No J,_, coupling is ever resolved.
1000 Hz
i-
Exercises 353
F F
Me2N-P
-- 81o
-100"
a. Explain the low temperature spectrum.
b. Explain the high temperature spectrum.
c. On the basis of these spectra, what can you say about the mechanism of exchange?
d. What is the relationship between coupling constants in the slow and fast exchange
regions?
5. The complex [(CF 3 )Co(CO) 3 (PF 3 )] [J. Amer. Chem. Soc., 91, 526 (1969)] is assumed
to be trigonal bipyramidal. The 19F nmr spectra at +30 C and -70 C are given
below. Splittings from "Co are not observed.
J =3Hz 9Hz 9Hz 3Hz 9Hz 3Hz
I70"C
0
6Hz 6Hz
[IIiA I III II, 30*C

500 0 500 1000 1500 Hz
Ref. H
CF3 PF3
First consider the -70 C spectrum.
a. Explain the reason why four quartets are observed for the fluorines bonded to
phosphorus. (No 59Co spin-spin coupling is detected.)
b. Explain how the CF 3 resonances are consistent with your explanation in part a.
c. Describe the reason for the smaller number of PF 3 resonances at 30 C.
d. What is the significance of the PF 3 resonances being observed as quartets in the

30 C spectrum?
6. The line width of the methyl proton resonance of 3-picoline-N-oxide in solutions

containing (3-picoline-N-oxide),Ni and excess ligand has been studied as a function
of temperature [Inorg. Chem., 10, 1212 (1971)]. The following plot was obtained.
2.0- 0
0
0
S1.5
00
.0000000
- so
1.0
o
0.5
2.5 3.0 3.5 4.0 4.5 5.0 5.5

103/T
a. Label the fast, near fast, intermediate, and slow exchange regions. Do this by giving
approximate boundaries to the region in units of 103 /T; e.g., fast: x, to x 2 units.
b. These data were analyzed using full line shape analysis from the classical treatment.
Why wasn't pyridine-N-oxide used as the ligand?
7. a. When the nmr spectrum of benzonitrile in a liquid crystal is observed, why do

solvent resonances not appear?
b. Why are the benzonitrile lines in a liquid crystal much sharper than those in solid
benzonitrile?
c. Compared to an nmr spectrum of benzonitrile in CCl4 , why are so many lines

observed in the liquid crystal nmr spectrum of benzonitrile?
8. The spectrum and structure of pyrogallol is given below. The OH protons are not
H
HO H
- HO H
OH
8.0 7.0 6.0 5.0 4.0 3.0 2.0 1.0 0Pm

Exercises 355
coupled to any other nuclei. Why don't the three phenyl protons give rise to a doublet
and a triplet? (The group of peaks centered about 6.7 ppm are expanded in the offset
sweep.)
9. 57Fe has an extremely small magnetic moment, and its abundance is only 2.2%. The
magnetic moment is not known to very good accuracy, and almost no information
exists about chemical shifts of 57 Fe nuclei. Even though the resonance signal was
expected to cover only a very small frequency range, and thus a slow passage scan
through the absorption signal would not take very much longer to observe than would
a free induction decay signal, this experiment was not reported until Fourier transform
nmr became available.
a. Why was Fourier transform nmr advantageous for this experiment?
b. Why wasn't any 57Fe-"C spin-spin splitting observed in the natural-abundance

sample?
10. Suppose that a sample has two sites with different chemical shifts. In the continuous
wave (CW) experiment, we expect a separate signal for each site.
2 2
T2 T2 8
WA WB
Consider a kinetic process that sets in as the temperature is raised, which exchanges
the nuclei between the two sites A and B. (In general, the problem is complex, since
the differential equations, one for each site, are coupled by the dynamic process and
the individual component magnetizations begin to lose their identities.) Consider the
limiting case in which the kinetic process begins to have an observable effect, yet one
can still consider the individual MA magnetization. Argue on a physical basis what
effect exchange will have on the width of the A-type resonance. (Hint: Recall the
physical process parametrized by T2A).
11. a. Let the resonance frequency of a water sample be coH2 0. Show how one can
determine the pulse duration T and a field strength H, for an oscillator tuned to
WH20 that will invert the H 2 0 magnetization.
b. Suppose that at a later time a 37/2 pulse was applied and that no free induction
decay curve was observed. What explains this? Of what use is this experiment?
c. A solute HX is added to the sample, and proton exchange occurs at a rate such
that separate but broadened resonances are observed for H 2 0 and HX Discuss
what effect is expected on solute signal intensity observed in a CW experiment if,
while scanning through it, one strongly irradiates the sample of WH20, saturating
the H 2 0 spin system.
12. In the experiment described in the text for measuring T1 , why must the 180 pulse
be followed by a 90 pulse?
13. In the slow passage nmr experiment, one frequently observes the following type of
pattern when sweeping the field from left to right:
The wiggles to the right of the absorption peak are referred to as "ringing" and are
found only when the nucleus observed has a long T2 . Explain ringing in terms of the
behavior of the bulk magnetization of the sample in this experiment. Be sure to specify
your frame of reference.
14. In the section on the effects of nuclei with quadrupole moments, there was a discussion
of the width of the 9 F resonance in the spectrum of NF 3 as a function of temperature.
a. Explain the quadrupole moment relaxation in terms of the concept of spectral

density.
b. Explain the resulting nmr spectrum in terms of spectral density.
15. Consider a frame of reference rotating at a frequency w, = 1.0000 x 108 Hz and a

sample of nuclei with a Larmor frequency of w, = 1.0001 x 108 Hz, with T = 100
sec and T2 = 0.01 sec. At t = 0, a 900 pulse is applied along the u axis using a strong
r.f. field of frequency w, in a negligible length of time. Using a uz or vs axis system
in each case, draw the net magnetization vector
a. just before the pulse.
b. just after the pulse.
c. at t = 2.5 x 10 sec.
d. at t = 1 sec.
e. at t = 104 sec.
16. The following 13 C T1's have been determined by Freemen [J. Chem. Phys., 54, 3367
(1971)] on 3,5-dimethylcyclohex-2-ene- 1-one:
Exercises 357
T (sec)
C1 37
0
C3 33
H
3-Me 5.9
H 6 2
C2 5.4
H 5 3 C5 5.3
4
C6 3.1
CH 3 Cl
C4 3.1
H H
5-Me 2.7
3
a. What appears to be the dominant means of spin-lattice relaxation for these C
atoms?
b. In light of your explanation, why do C, and C 3 have such long T,'s?
17. Experiments can be performed to determine the T, of a specific atom in a molecule.

The most commonly employed technique using FTNMR is to apply a 180 pulse,
delay for a time T, followed by a 90 pulse and immediate acquisition. Tau is usually
varied from 0 see to many times the expected T value.
a. Sketch a generic version of this pulse sequence showing the bulk magnetization after
each step in the experiment.
b. Consider the magnetization following the 90 pulse previously stated. Show what
the spectral peaks (assume a singlet peak) would look like at the following times:
T = 0, T = fraction of T1, T = T, T > T1.
c. Would it be possible to perform this experiment to measure T, using only a 90

pulse, waiting T, and then acquiring? Explain.
18. The proton-coupled "B COSY spectrum of decaborane is shown as follows.
a. Draw the "B proton-decoupled nmr spectrum.
b. Assuming only directly bound borons are coupled, assign the resonances in the
"B nmr spectrum. Explain.
So
I I I I I
20 10 0 -10 -20 -30
S("B)/ppm
19. The 'H COSY spectrum of a sample is shown below (letters correspond to the same
"C nucleus throughout this problem).
4
E
a.
5>
cco
10 9 8 7 6 5 4 3
8('H)/ppm
a. Draw the proton spectrum.
b. Using the letters used to label carbons, assign the coupled peaks by letter, e.g., A
is coupled to B and C.
The 'H/' 3 C HETCOR spectrum of the sample above is shown next.
B C G FA H
I I I I I I I I 0
0 -1
2
9
3
4 E
I 5>
9
6
S -7
I I I I I
- -
10
150 100 50
S(1Cl/ppm
c. Draw the proton-decoupled 13 C spectrum.
13
The heteronuclear C 2-D, J-spectrum is shown as follows.
D B C E G FA H
---- . . i200
100
ME 9
ME I
-100
I I I I I I I I I I I
50 -200
S( C)/ppm
d. Describe how the J-spectra aid in peak assignment.
e. The molecule is a partially unsaturated ester with empirical formula C8 H 40 2.

Draw the structure.
20. The following spectra (acidic conditions) were taken for the molecule whose structure
is given below. The J-spectrum is given in A and the HETCOR in B. Assign the
numbered peaks to the labeled carbons (ignore all unlabeled carbons).
-TMv S - 0
OH CH 3
b 5
HC3_
O a b c 10
E
200 (2)
'CH32
15
(3)
(4) F2 20
(5)
(3)
(4) - 25 0
--- (5)
I -I
30
( -
(6) - 35
'
i 4 3 2 1 0- -1 -2
0 Hz +100
Proton Shifts (ppm)
Electron Paramagnetic
Resonance
Spectroscopy*
Introduction 9-1 PRINCIPLES
Electron paramagnetic resonance is a branch of spectroscopy in which radiation

of microwave frequency is absorbed by molecules, ions, or atoms possessing
electrons with unpaired spins. This phenomenon has been designated by different
names: "electron paramagnetic resonance" (epr), "electron spin resonance" (esr),
and "electron magnetic resonance" (emr). These are equivalent and merely
emphasize different aspects of the same phenomenon. There are some similarities
between nmr and epr spectroscopy that are of help in understanding epr. In nmr
spectroscopy, the two different energy states (when I = 1/2) arise from the alignment
of the nuclear magnetic moments relative to the applied field, and a transition
between them occurs upon the application of radio-frequency radiation. In epr,
different energy states arise from the interaction of the unpaired electron spin
moment (given by m, = ± '/2) with the magnetic field, the so-called electronic
Zeeman effect. The Zeeman Hamiltonian for the interaction of an electron with
the magnetic field is given by equation (9-1):
kI= g#HS (9-1)
where g for a free electron has the value 2.0023193, #l is the electron Bohr
magneton, eh/2mec, which has the value 9.274096 + (0.000050) x 10-2 erg
gauss' 5, is the spin operator; and H is the applied field strength. This
Hamiltonian operating on the electron spin functions a and #l corresponding to
m, = + '/2 and '/2, respectively, produces the result illustrated in Fig. 9-1. The
# spin state has its moment aligned with the field, in contrast to nmr, where the
lowest energy state corresponds to m, = + 1/2(oN). The lowest energy state in
epr corresponds to m, = - '/2 because the sign of the charge on the electron is
opposite that on the proton.The transition energy is given by equation (9-2):
AE = g#H (9-2)
* The Additional References contain reviews of epr. The reference by J. E. Wertz and J. R. Bolton
360 is especially recommended.
9-1 Principles 361
E = +(112)g3H
FIGURE 9-1 The removal of

am, = -1 /2)*
the degeneracy of the a and
# electron spin states by a
- 1/) magnetic field. *Note the
difference in the ground state
from nmr.
H = 0 H inc.-
The energy difference between the a and fl spin states in magnetic fields of strengths
commonly used in the epr experiment (several thousand gauss) corresponds to
frequencies in the microwave region. For a field strength of 10,000 gauss, one
can calculate from equation (9-2) that AE is 28,026 MHz. This is to be contrasted
with the energy for the transition of the proton nuclear moment of 42.58 MHz
(i.e., radiation in the radio-frequency region) in a magnetic field of identical
strength. The magnetic moment for the electron is -- 9.2849 x 10 -" erg gauss 1
compared to 1.4106 x 10-3 erg guass-1 for a proton nuclear moment.
The epr experiment is generally carried out at a fixed frequency. Two common
frequencies are in the X-band frequency range (about 9500 MHz or 9.5 gigahertz,
GHz,* where a field strength of about 3400 gauss is employed) and the so-called
Q-band frequency (35 GHz, where a field strength of about 12,500 gauss is used).
Since the sensitivity of the instrument increases roughly as v2 and better spectral
resolution also results, the higher frequency is to be preferred. There are several
limitations on the use of the Q-band frequency. Smaller samples are required for
Q-band, so the sensitivity is not as much greater as one would predict from v2 .
It is more difficult to attain the higher field homogeneity (6H/H) that is required
at higher frequencies. Finally, for aqueous samples, dielectric absorption by the
solvent becomes more serious as the frequency increases and this results in
decreased sensitivity.
Water, alcohols, and other high dielectric constant solvents are not the
solvents of choice for epr because they strongly absorb microwave power. They
can be used when the sample has a strong resonance and is contained in a
specially designed cell (a very narrow sample tube). Epr measurements on gases,
solutions, powders, single crystals, and frozen solutions can be carried out. The
best frozen solution results are obtained when the solvent freezes to form a glass.
Symmetrical molecules or those that hydrogen bond extensively often do not
form good glasses. For example, cyclohexane does not form a good glass, but
methylcyclohexane does. Some solvents and mixtures that form good glasses are
listed in Table 9-1.
The sample tube employed is also important. If the signal-to-noise ratio is
low, a quartz sample tube is preferred, because Pyrex absorbs more of the
microwave power and also exhibits an epr signal.
* The prefix mega indicates 10', and giga indicates 109.

362 Chapter 9 Electron Paramagnetic Resonance Spectroscopy
TABLE 9-1. Commonly Used Glasses*

Pure Substances
3-Methylpentane Sulfuric acid Sugar

Methylcyclopentane Phosphoric acid Triethanolamine
Nujol (paraffin oil) Ethanol 2-Methyltetrahydrofuran
Isopentane Isopropanol Di-n-propyl ether
Methylcyclohexane I-Propanol cis-trans Decalin
Isooctane -Butanol Triacetin
Boric acid Glycerol Toluene
Mixtures
Components Ratio A/B
3-Methylpentane/isopentane 11
Isopentane/methylcycloheXane 16
Methylcylopentane/methylcYclohexane 1/1
Pentene-2(cis)/pentene- 2 (trals)
Propane/propene 11
Isopropyl benzene/propane/propene 2/99
Ethanol/methanol 4/1, 52, 19
Isopropyl alcohol/isopentane 37
Ethanol/isopentane/diethyl ether 2 5t5
Alphanol 79'/mixture of primary alcohols
Isopentane/n-butyl alcohol 7 D3
Isopentane/isopropyl alcohol 8/2
Isopentane/n-propyl alcohol 82
Diethyl ether/isooctanelisopropyl alcohol 31311
Diethyl ether/isooctanenethyl alcohol 3A3/1
Diethyl ether/isopropyl alcohol 311
Diethyl ether/ethanol 3/1
Isooctane/methylcyclohexane/isopropyl alcohol 331/
Diethyl ether/toluene/ethanol 21/1
Isopropyl alcohol/isopentane 2/5
Propanol/diethyl ether 2/5
Butanol/diethyl ether 25
Diethyl ether/isopentane/dimethyl formamide3ethanol 12/106/1
Water/propylene glycol 1/1
Ethylene glycol/water 2/1
Trimethylamine/isopentane/diethyl ether 2/5/5
Triethylamine/isopentanediethy ether 3/1/3
Methylhydrazine/methylaminetrimethylamine 24/2
Diethyl ether/isopentane ethanol/pyridine 12/10/6/1
Di-n-butyl ether/diisopropyl etheradimethyl ether 3/5/12
Diphenyl ether/o-diphenyleth anetriphenylmethane 3/3/1
Diethyl ether/isopentane 11 to 1 2
Dipropyl ether/isopentane 31
Dipropyl ether/methylcylohexane 3/1
Diethyl ether/pentene-2(cis)-pentene-2(trans) 2/1
Ethyl iodide/isopentane/diethyl ether 12/2
Ethylbromide/met hylcyclohexane/iso pen tane/methylcycl opent ane 1/4/7/7
Ethanol/methanol/ethyl iodide 16/4/1
Ethanol/methanol/propyl iodide 16/541
Ethanol/methanol/propyl chloride 16/4/1
Ethanolmethanol/propyl bromide 16/4/1
Diethyl ether/isopentane/ethanol/l-chloronaphthalene 862/2
3-Methylpentane/isopentane 1/2
Propyl alcohol/propanepropene 2/4/9
Diisopropylamine/propane/propene 2/9/9
Dipropyl ether/propane/propene 2/44
Mixtures
Components Ratio A/B
Toluene/methylene chloride / or excess toluene

Toluene/acetone 1/1 or excess toluene
Toluene/methanol or ethanol 1/1 or excess toluene
Toluene/acetronitrile 1/1 or excess toluene
Toluene/chloroform 1/1 or excess toluene
2-Methyltetrahydrofuran/methanol 2/1
2-Methyltetrahydrofuran/proprionitrile 2/1
2-Methyltetrahydrofuran/methylene chloride 1/1
*Abstracted in part, with permission, from B. Meyer, "Low Temperature Spectroscopy,"
American Elsevier Publishing Co., New York, 1971
There are many effects that modify the electron energy states in a magnetic
field. We shall consider these factors one at a time by discussing the epr spectra
of increasingly complex systems.
In the way of introduction, it will simply be mentioned that differences in
the energy of the epr transition for different molecules are described by changing
the value of g in equation (9-2). This is to be contrasted with nmr, where one
customarily holds gN fixed and introduces the shielding constant to describe the
different resonance energies, i.e., for nmr,
AE - -gNN(l - a)HAm1, (9-3)
As we proceed to more complex systems, we shall discuss the factors that influence
the magnitude of g.
9-2 THE HYDROGEN ATOM Nuclear

The first contribution to epr transition energies that will be introduced is the
Hyperfine
electron-nuclear hyperfine interaction. The hydrogen atom (in free space) is a Splitting
simple system to discuss because, by virtue of its spherical symmetry, anisotropic
effects are absent. In the development of epr, we shall employ the Hamiltonian
to quantitatively describe the effects being considered. The full interpretation of
the esr spectrum of a system is given in terms of an effective spin Hamiltonian.
This is a Hamiltonian that contains those effects, of the many to be described,
which are used to interpret the particular spectrum of the compound studied.
The complete spin Hamiltonian for the hydrogen atom (in free space) is
A = gH -$ - gN3NH-Z + aZ
= gf(H S + HS,+ H S.) -gNN(Hxx + Hy + Hz) + aZ,$,
+ aIs, + aZ,$A
For a spherical system in a magnetic field that is defined as the z-axis, this
simplifies* to
N = gf3Hz - gNOfNHIz + aZ-S (9-4)
The first term of this Hamiltonian discussed earlier [equation (9-1)], leads to the
energy-field relation shown in Fig. 9-1. The second term of the Hamiltonian is
familiar from our discussion of nmr; it describes the interaction of the nuclear
moment of the hydrogen atom with a magnetic field. It is of opposite sign (the
state with m, = + 1/2 is lowest) and smaller in magnitude than the first term.
The combined effect of the first two terms in equation (9-4) upon the energies
of the spin states of the hydrogen atom in a magnetic field is shown in Fig. 9-2(C).
The field strength is fixed in this figure, and the dashed lines simply show the
energy changes incurred by adding a new term in the Hamiltonian. In order to
determine the energy of the hydrogen atom in a magnetic field, we employ a
basis set for this Hamiltonian [equation (9-4)] that consists of the four possible
electron and nuclear spin functions. Such a basis set is (p, = |e N>
, 2 e NflN>,
(p3 = I0e N> and (P4 = Ie N>. Let us begin by calculating the energies arising
from the first two terms in the Hamiltonian, ho. We must solve the simultaneous
equations <(pH I I ,> - E<p I Pm> = 0, where n and m may or may not be
equal. Thus the 4 x 4 secular determinant in this basis set contains diagonal
terms of the type:
eN IgfzH. - gNNHI 1efN - EOgCN eN
= eNINgflHgzI e> IN> - KXN INI NHIZ IN e)10- E=0
Since for this problem the operators are Iz and S., a and # are eigenfunctions;
i.e., Sz 1 e) = '/2e, z lOe = - '20e, 2 |)N 2aN and z flN> - 2#N'
Furthermore, S. does not operate on the nuclear spin function and Z does not
operate on the electron spin function, leading to:
2 2
It should now be clear that all the off-diagonal elements will be zero with this
Hamiltonian, for all of the off-diagonal elements are of the form
<(p, Ho (p.> - (P (Pm>, which equals zero when n = m. Since the Hamiltonian
matrix is diagonal, the determinant is already factored and we get the four energies
directly, as shown above for Je#N' They are indicated for E,, E2 , E3 , and E 4 , in
Fig 9-2(C). These results can be verified for practice. The usual selection rules
in epr are Am, = 0 and Am. = + 1. It will be noticed that the two esr transitions
(Am, = 0) illustrated in Fig 9-2(C) have the same energy. Considering only the
first two terms of the Hamiltonian, the epr spectrum of the hydrogen atom would
be the same as that of a free electron, i.e., one line at a field hv/gf#or g = 2.0023.
Next, we must concern ourselves with the a! - 5 term in the Hamiltonian. This
term describes the coupling of the electron and nuclear spin moments, which
* H_ and H, are zero and H = H. The effects of S,, S,, I., and I, are not necessarily zero.
-L---ZE
Iae ON ) - --
~,-1
]aOON)
he EF ---
N
/ 2 'N
'N
N
Am 2
U/
p~N
OeN
-- E4
gH
- E3-
gPH~~-
e N
3
- 9NO
3
E I=
ge- gN NH
2
a
4(gOH +gN ONH)
E4=-jgH +9N9NH +0
-a
E =+jgLH-jgNNH +a +0
2
E + jg3H +
2
E2=+2gpH
Igf~
2NONH + a
4(g3H +gN ONH)
(A) (B) (C) (D) (E)
Free Electron, H= goHS, H =gpHS- Hi,
gN1ON H =gpHS, -NONH/z H = gHS
2 - H/
No Magnetic Field
+ a/, ' Sz + az ',
FIGURE 9-2 The influence of various terms in the Hamiltonian on the energy of a hydrogen atom in a magnetic
field. #eaN
corresponds to ms = - 1/2 and m, = +1/2; feflN to ms = 1/2 and m, = 1/2; mef#N to mS = +1/2 and m, 1/2; and aeON to
ms = + 1/2 and m, = + 1/2.
classically corresponds to the dot product of these two vectors. The quantity a
indicates the magnitude of the interaction and has the dimensions of energy. This
is referred to as the Fermi contact contribution to the coupling, and its magnitude
depends upon the amount of electron density at the nucleus, 0o)2, according to:
a 8=ygpNN
83 (0)I2 (9-5)
For a hydrogen atom, the Slater orbital function is 0, = (1/mao 3 ) exp (-r/a,)
where ao, the Bohr radius, equals h2 /me 2 = 0.52918 A. Substituting into equation
(9-5), the value of $ for a hydrogen s-orbital at r = 0 yields a/h = 1422.74 MHz.
Since the nuclear hyperfine interaction we are talking about involves the dot
product of the nuclear and spin moments, it has x-, y-, and z-components, so
aZ -$ = a(Z,$x + Z$ + Z$) (9-6)
The elements generated by the 1z3z term operating on the basis 2,2N, etc., are
again only diagonal elements because (<p. I aZ $SI p.> = 0 when m # n; for
example, <a,#N I aI aN> = 0. The following results are obtained for the diagonal
elements:
(a~eoN I eaS.I - a
~V(N>
4
06A3I a3.1I 'A3> = - 1a

41
(pexN I a3.1 I3e N> a
(ffe/N Iazj. I e#N> a
These have to be added to the energies in Fig. 9 2(C), modifying the energies as
shown in Fig. 9-2(D). The contributions of +('/ 4 )a to Ei, E2 , etc., from afzgz are
indicated at the bottom of (D). Now we see, looking at the arrows for the two
electron spin changes, that the transition energies are no longer equal. One
transition gives rise to a spectral peak at lower energy than that corresponding
to g = 2.0023 [see Fig. 9-2(D)] by ( 1/2 )a, and the other occurs at an energy that
is higher by ('/ 2 )a. The energy separation of the two peaks is a.
To complete the problem, we now have to add the effects of 7N$, and I,5,.
This is best done in terms of the raising and lowering operators, which work in
a similar fashion to !+ and I that were discussed earlier. For the electron spin
operators, we define:
S= Sx + iS
S = S i
Thus
$+Z = ($ Z + $,Z,) + i($ - $,)

and
$ Z $Zj + $,Z, -- i(, - 3jZ).
Combining these equations, we see that:
"L~+S~~ -Y ±2i
The following results are obtained by analogy to our earlier discussion of 1+ and
I_:
S_ I+I e/N>=IflXN>
MS aIOeN>= 0
All other operations of $_Z± or 5SZ_ upon the basis set produce zero. Thus, if
we consider the 4 x 4 matrix shown in Fig. 9-3, the only non-vanishing matrix
elements from g f_ and g- I+ are
('eNIaS+ flON> =a
(#CN Ia - + e4N a
IaeaN> IajN> Ie&No IfleN>
lae/3N)
I M10~
|fenN)
|fiefNO
FIGURE 9-3 Secular determinant for the field-free hydrogen atom.
The matrix element
1
( cief0N xSIx ySIy O~ecN
1
(exCN xaSjl + aSyi, 140
e = a
We can summarize this entire section by completing the full determinant for
the original spin Hamiltonian equation (9-4), operating on the p basis set to give
energies <(P, I R IH = E(p. p,,>. The determinant shown in Fig. 9-3 equals
zero. Note that it is block diagonal so that two of the energies, E, and E4 , are
obtained directly. We also see that !,, and 7,$, lead to off-diagonal elements
that mix 92 and 93. A perturbation theory solution* of the resulting 2 x 2
determinant gives (to second order):
E 1 3 H 1 a H 1 a2
E2 =-Ig#H +-g NNH a+ a
2 2 4 4(gflH + gNNH)
E3 =- -g#H+-gN0NH) a-
2 2 4 4(gplH + gN#NH)
We can see in Fig. 9-2(E), where the effects of these off-diagonal elements on the
energy levels are illustrated, that the energies of both transitions are increased
by the same amount. Since the off-diagonal elements are small compared to the
diagonal, effects arising from this part of the Hamiltonian are referred to as
second-order effects. Second-order effects thus have no influence on the value of
a read off the spectrum, but will change the value read off the spectrum for g. A
more interesting contribution to the spectral appearance is that the previously
forbidden transition E3 -- E2 (the simultaneous electron and nuclear spin flip),now
becomes allowed because of the mixing of the basis set.t
9-3 PRESENTATION OF THE SPECTRUM
As in nmr, the epr spectrum can be represented by plotting intensity, I, against

the strength of the applied field; but epr spectra are commonly presented as
derivative curves, i.e., the first derivative (the slope) of the absorption curve is
plotted against the strength of the magnetic field. It is easier to discern features
in a derivative presentation if the absorption lines are broad. The two modes of
presentation are easily interconverted, and the relationship between the two kinds
of spectra is illustrated in Fig. 9-4. In (A), a single absorption peak with no fine
structure is represented; (B) is the derivative curve that corresponds to (A). The
derivative curve crosses the abscissa at a maximum in the absorption curve, for
the slope changes sign at a maximum. Curve (C) is the absorption counterpart
of curve (D). Note that the shoulders in (C) never pass through a maximum and
as a result the derivative peaks in (D) corresponding to these shoulders do not
cross the abscissa. The number of peaks and shoulders in the absorption curve
can be determined from the number of minima (marked with an asterisk in (D))
or maxima in the derivative curve.
2
*The exact solution is E = -(%) ai (1/2) [(g# + gN N 2 2 a ]1 2
t This allowedness is derived in Carrington and McLachlan (see Additional References) on

page 401.
9-3 Presentation of the Spectrum 369
9995 10000 10005 gauss 9960 10000 10040 gauss

(A)
FIGURE 9-4 Comparison of
spectral presentation as
absorption (A and C) and
derivative (B and D) curves.
9995 10005 gauss gauss
The epr spectrum of the hydrogen atom is illustrated in Fig. 9-5. To a good
approximation, the g-value is measured at x, which is midway between the two
solid circles corresponding to absorption peak maxima. The hyperfine splitting,
a/gf#, is the separation between the solid circles in gauss. The sign of a generally
cannot be obtained directly from the spectrum. The splitting in Fig. 9-2 implies a
positive value for a. If a were negative then m, -1/2 and m,=-'/ 2; that is
(##N) would be the low energy state.
The epr spectrometer is designed to operate at a fixed microwave source
frequency. The magnetic field is swept, and the horizontal axis in Fig. 9-5 is in
units of gauss. One can set the field at any position using the field dial and sweep
from that place. For a fingerprint type of identification, greater accuracy is needed
than can be obtained using the instrument dials. For this purpose, an external
standard, diphenylpicrylhydrazide DPPH, is used together with a microwave
frequency counter. DPPH has a g-value of 2.0037 + 0.0002. The field sweep is
assumed to be linear, and the g values of other peaks are calculated relative to
FIGURE 9-5 The epr

spectrum of a hydrogen atom
(#l has units of ergs/gauss).
H (gauss)
this standard. The field axis is in units of gauss, and the g value is reported as
a dimensionless quantity using
hv
g =
#lH
where v is the fixed frequency of the probe and H (which is being swept) is
obtained from the spectrum. A frequency counter should be used to measure the
probe frequency v.
The value of a is sometimes reported in units of gauss, MHz, or cm-'. It is
to be emphasized that the line separation in a spectrum in units of gauss is given
by a/# (where a has units of ergs and # has units of ergs/gauss -'). When g #
2, it is incorrect to report this separation as a in units of gauss. One would have
to multiply the line separation by g#l and divide by g,# (where g, is the free
electron value of 2.0023193) to report a correct value for a in gauss. Since a is
an energy, it is best to report its value as an energy. This is simply done by
multiplying the line separation in gauss by g#3, with # in units of cm- /gauss-.
There is no g-value dependence for this unit. The value of a in MHz is obtained
by multiplying a (cm-1) by c (3 x 101 cm sec-1) and dividing by 106.
9-4 HYPERFINE SPLITTINGS IN ISOTROPIC SYSTEMS

INVOLVING MORE THAN ONE NUCLEUS
The first-order energies of the levels in the hydrogen atom are given by equation
(9-7), which ignores the small nuclear Zeeman interaction.
E = gpHm, + amsm, (9-7)
Substituting the values of m, and m, into this equation enables one to reproduce
the energies given in Fig. 9-2(D). For a nucleus with any nuclear spin, the
projection of the nuclear magnetic moment along the effective field direction at
the nucleus can take any of the 21 + I values corresponding to the quantum
numbers -I,(-I + 1),..., (I - 1), 1. These orientations give rise to 21 + 1
different nuclear energy states (one for every value of mi); and when each of these
couples with the electron moment, 21 + 1 lines result in the epr experiment. Since
these energy differences are small, all levels with the same m. value are equally
populated for practical purposes and the epr absorption lines will usually be of
equal intensity and equal spacing. For example, three lines are expected for an
l'1.1 NJ 1.14 unpaired electron on 4 N, whose I = 1.
We shall next consider the effect on the spectrum when the electron interacts
with (i.e. is delocalized onto) several nuclei. For simplicity, assume that the species
is rotating very rapidly in all directions, and that the g value is close to the free
144 3 electron value. As an example, the methyl radical will be discussed.!" As
illustrated in Fig. 9-6, addition of the nuclear spin angular momentum quantum
FIGURE 9-6 Possible numbers of the individual protons results in four different values for the total
nuclear spin arrangements of nuclear spin moment, Mr. As indicated in Fig. 9-7, this gives rise to four transitions
the protons in a methyl (AM, = 0, Am, = + 1). Since there are three different possible ways to obtain a
radical. total of M, = + /2 or -'/2 (see Fig. 9--6), but only one possible way to obtain
9-4 Hyperfine Splittings in Isotropic Systems involving More Than One Nucleus 371
2 FIGURE 9-7 The four transitions that

occur in the epr spectrum of the methyl
radical (see Fig. 9-8 for the spectrum).
(As with the H atom, the +m, state is
lowest for m, z - 1/ 2 and the -m, state
is lowest for m. = +1/2, from the f- $ term.)
3
2
2 +2
+3
2
M = +3/2 or 3/2, the former system is three times more probable than the
latter and the observed relative intensities for the corresponding transitions (Fig.
9-7) are in the ratio 1:3:3:1.
In general, when the absorption spectrum is split by n-equivalent nuclei of
equal spin Ii, the number of lines is given by 2n1 + 1. When the splitting is
caused by both a set of n-equivalent nuclei of spin I, and a set of m-equivalent
nuclei of spin Ii, the number of lines is given by (2n!I + 1)(2mI + 1). The following
specific cases illustrate the use of these general rules.
1. If a radical contains n non-equivalent protons onto which the electron
is delocalized, a spectrum consisting of 2" lines will arise.
2. If the odd electron is delocalized over a number, n, of equivalent protons,
a total of n + I lines, (2nI + 1), will appear in the spectrum. This number is less
than the number of lines expected for non-equivalent protons (i.e., 2") because
several of the possible arrangements of the nuclear spins are degenerate (see Fig.
9-6). The spectrum of the methyl radical illustrated in Fig. 9-8 contains the four
peaks expected from these rules.
The spectra expected for differing numbers of equivalent protons can easily
be predicted by considering the splitting due to each proton in turn, as illustrated
in Fig. 9-9. When the signal is split by two equivalent protons, the total M, for
the three levels can have the values Em, = + 1, 0, and -1. Since there are two
ways in which we can arrange the separate m1 's to give a total M, = 0
(namely + 2', - 2, and -1/2, +'/2),the center level is doubly degenerate. Three
peaks are observed in the spectrum (AM, = 0, Am, = i 1), and the intensity ratio
is 1: 2: 1. The case of three protons (e.g., the methyl radical) was discussed above,
and similar considerations are employed for the systems represented in Fig. 9-9 Inc. field strength
with more than three protons. FIGURE 9-8 The derivative
The relative intensities of the peaks are given by the coefficients of the spectrum of the methyl
binomial expansion. It should be remembered in applying this formula that it is radical in a CH 4 matrix at 4.2
restricted to equivalent protons or other nuclei having I = /2- K.
FIGURE 9-9 Hyperfine

energy levels resulting from
interaction of an unpaired
electron with varying
numbers of equivalent
protons 2( Each number in
parentheses gives the
degeneracy of the level to
which it refers, and hence
the relative peak intensities
for the corresponding
transitions.
0 1 2 3 4 5
Number of protons
3. If the odd electron is delocalized over two sets of non-equivalent protons,

the number of lines expected is the product of the number expected for each set
[(2nI; + 1)(2mI + 1)]. The naphthalene negative ion, which can be prepared by
adding sodium to naphthalene, contains an odd electron that is delocalized over
the entire naphthalene ring. Naphthalene contains two different sets of four
equivalent protons. A total of n + 1, or five peaks, is expected for an electron
delocalized on either set of four equivalent protons. In the naphthalene negative
ion, the two sets of four equivalent protons should give a total of 25 lines in the
epr spectrum. This is found experimentally.
4. If the electron is delocalized on nuclei with spin greater than %/2,
a procedure similar to that for protons can be applied to calculate the number
of peaks expected. If the electron is delocalized over several equivalent nuclei
that have spins greater than ', the number of peaks expected in the spectrum
is predicted from the formula 2nI + 1. For example, five peaks are expected for
an electron delocalized on two equivalent nitrogen atoms. A procedure similar
to that in Fig. 9-6 shows that the intensities of the five peaks will be in the ratio
a 1:2:3:2:1.
5. If the electron is delocalized over several non-equivalent atoms, the
I I I (A) total number of peaks expected is obtained by taking the product of the number
expected for each atom. The scheme illustrated in Fig. 9-10 for an electron
delocalized onto two non-equivalent nuclei with I = I is often employed to
111 Ill I11 (B) indicate the splitting expected. The three lines in (A) represent splitting of an epr
peak by a nucleus with I = 1 and a hyperfine coupling constant a. Each of these
a lines is split into three components as a result of delocalization of the electron
on a second, non-equivalent nucleus with I = 1, and a hyperfine coupling constant
FIGURE 9-10 (A) Three
a', producing in (B) a total of nine lines. In subsequent discussions a scheme
lines expected from an
electron on a nucleus with similar to that in Fig. 9-10 will be employed for the interpretation of spectra.
/ = 1, and (B) nine lines The shape of the spectrum and the separations of the peaks will depend upon
resulting from the splitting by the resonant field, g, and the coupling constants a and a'. Frequently the measured
a second non-equivalent spectrum will not reveal all the lines expected, because the line widths are large
nucleus with /= 1. compared to a/g# and two close lines are not resolved. For example, the spectrum
9-4 Hyperfine Splittings in Isotropic Systems Involving More Than One Nucleus 373
in Fig. 9-11 could result for the hypothetical radical H-( + <- -X + where
I = 1 for X. The two lines in (A) result from the proton splitting. In (B) each line
in turn is split into three components owing to interaction with nucleus X; thus
we would expect six lines, all of equal intensity. However, it is possible to detect
only five lines, if the two innermost components are not resolved. They would
give rise to a single peak with twice the area of the other peaks (see Fig. 9-11). (A) I I
The epr spectrum 3 ) of bis-salicylaldimine copper(II) in Fig. 9-12 is an (B) I I I I I
interesting example to summarize this discussion of nuclear coupling. This FIGURE 9-11 Hypothetical
spectrum was obtained on a solid and is not isotropic; this aspect will be discussed absorption spectrum for the
shortly. Four main groups of lines result from coupling of the 6 3Cu nucleus radical H -X+ (/ = 1 for X).
(I = 3/2) with the electron. The hyperfine structure in each of the four groups
consists of eleven peaks of intensity ratio 1:2:3:4:5:6:5:4:3:2:1. These
peaks result from splitting by the two equivalent nitrogens and two hydrogens,
H' in Fig. 9-12. The total number of peaks expected is fifteen;
(2 nNIN + 2
nH'HH)( + 1) = 5 x 3 = 15. The eleven peaks found for each subgroup
in the actual spectrum result from overlap of some of the fifteen peaks as indicated
in Fig. 9-13. The line for an electron not split by a nucleus is shown in (A). The
splittings by the two equivalent nitrogens are indicated in (B) and the subsequent
splitting by two equivalent protons is indicated in (C) The two nitrogens split
the resonance into five peaks of relative intensity 1:2:3:2:1. These values are
denoted in (B) by 4d, 4e, and 4f where the intensities correspond to d = 1, e= 2,
and f= 3. The splitting by two equivalent protons will give rise to three lines
for each line in (B), with an intensity ratio of 1:2: 1. The intensities indicated by
letters underneath the lines in (C) result from the summation of the expected
intensities. Since the relative intensities are d = 1, e = 2, and f= 3, the ratio of
the intensity of the bands in (C) is 1:2:3:4:5:6:5:4:3:2:1. The experimental
spectra agree with this interpretation, which is further substantiated by the
following results:
1. Deuteration of the N-H" groups (see Fig. 9-12) produced a compound
which gave an identical spectrum.
2. When the H' hydrogens were replaced by methyl groups, the epr spectrum
for this compound consisted of four main groups, each of which consisted of five
4 FIGURE 9-12
EPR-derivative spectrum of
bis-salicylaldimine copper(Il)
with isotopically pure "Cu.
Askerisk indicates calibration
peak from DPPH. From A. H.
Maki and B. R. McGarvey, J.
2 H]2 Chem. Phys., 29, 35 (1958).
Decreasing H
FIGURE 9-13 Interpretation

of the epr spectrum of I I
bissalicylaldimine copper(ll). 4e 4d
(A) An unsplit transition. (B)
Splitting by two equivalent (C) I I II I II I II I II I I
nitrogens, with d, e, and f d, 2d, d + e, 2e, e + f,2f, f + e, 2e, e + d, 2d, d
indicating relative intensities
of 1, 2, and 3. (C) Further
splitting by two equivalent lines resulting from nitrogen splitting only. The hyperfine splitting by the N-H"
protons. proton and that by the protons on the methyl group are too small to be detected.
This spectrum furnishes conclusive proof of the delocalization of the odd
electron in this complex onto the ligand. This can be interpreted only as covalence
in the metal-ligand interaction, for only by mixing the metal ion and ligand wave
functions can we get ligand contributions to the molecular orbital in the complex
that contains the unpaired electron.
FIGURE 9-14 (A) Basic Another interesting application of epr(4 ) involves the spectrum of
molecular geometry of the Co 3 (CO),Se, whose structure and epr spectrum are shown in Fig. 9-14. The
Co3(CO) 9Se complex. (B) The 22-line spectrum indicates that the one unpaired electron in this system is
epr spectrum of a single completely delocalized over the three cobalt atoms (Ic = /2). This in effect gives
crystal of FeCo 2(CO),Se rise to an oxidation state of +2/3 for each cobalt atom.
doped with about 0.5% of
paramagnetic Co 3(CO),Se.
This spectrum containing 22
hyperfine components was
recorded at 77 K with the
molecular threefold axis
parallel to the magnetic field
direction. [Reprinted with
permission from C. E.
Strouse and L. F. Dahl, J. (A) (B)
Amer. Chem. Soc., 93, 6032
(1971). Copyright by the
American Chemical Society.]
9-5 CONTRIBUTIONS TO THE HYPERFINE

COUPLING CONSTANT IN ISOTROPIC SYSTEMS
In Equation (9-5), we saw that:

8ir
a = 37 9#gN#N 10012
In a molecule, the hyperfine splitting from delocalization of unpaired spin density

p(rN) onto a hydrogen atom of the molecule is given by equation (9-8):
-8nr
a = 8ggN#NpH(rN) (9-8)
where p(rN) can be crudely thought of as the difference in the average numbers
of electrons at the nucleus that have spin moments characterized by m, = + 1/2
and - 1/2. When there is an excess of + '/2 spin density (i.e., the electron spin
9-5 Contributions to the Hyperfine Coupling Constant in Isotropic Systems 375
moment opposed to the field) at the nucleus, this nucleus is said to experience
negative spin density. An excess of electron density with m, = - 1/2 is called positive
spin density. Positive spin density is represented by an arrow that is aligned with
the external field, and negative spin density is represented by one opposed to the
field. To further complicate matters, positive spin density is often referred to as
a spin in the literature, even though the wave function for evaluation of the matrix
elements for this electron is represented by fl. Thus, the common convention
for labeling spin density is exactly the opposite of that used to label the electron
spin wave functions. We will avoid the a and # labels of spin density and use
these as symbols for the spin wave functions in this book.
We should also be careful to point out that the amount of unpaired spin
density on an atom in the molecule does not correspond directly to the atom
contributions in the molecular orbital containing the unpaired electron. We shall
refer to the latter effect as unpaired electron density. An unpaired electron in an
orbital of one atom in a molecule can polarize the paired spins in an orthogonal
sigma bond so that one of the electrons is more often in the vicinity of one atom
than in the vicinity of the other. This puts unpaired spin density at the nucleus
of the atom even though there is no unpaired electron density delocalized onto
it. We can make this more specific with the following example.
Some of the first attempts at interpreting hyperfine couplings involved
aromatic radicals with the unpaired spin in the 7n-system, e.g., CHNO2 . Hckel
calculations were carried out and the squares of the various carbon p. coefficients
in the m.o. containing the unpaired electron were employed to give the amount
of unpaired electron density on the various carbon atoms. The hyperfine splittings
observed experimentally were from the ring hydrogens, which are orthogonal to
the nt-system. No unpaired electron density can be delocalized directly onto them,
but unpaired spin density is felt at the hydrogen nucleus by the so-called
spin-polarization or indirect mechanism. We shall attempt to give a simplified
view of this effect using a valence bond formalism. Consider the two resonance
forms in Fig. 9-15 for a C-H bond in a system with an unpaired electron in a
carbon p. orbital. In the absence of any interaction between the 7E and a systems,
the so-called perfect pairing approximation, we can write a valence bond
description of the bonding and antibonding sigma orbital wave functions:
11
($' + #n) and * -'f)
2 (0 ,2
Here f, and 0, represent wave functions for structures I and II in Fig. 9-15,
which we shall not attempt to specify in terms of an a.o. and spin basis set. When
interaction of the 7t and a systems is considered, we find that I is a more stable
structure than II and, accordingly, it contributes to the ground state more than
does II, resulting in valence bond functions:
0' ap, + boll (a > b)
and FIGURE 9-15 Resonance

forms for a C-H sigma bond
fragment with an unpaired
* a'f, - b'oll (a' < b') electron on the carbon.
This in effect polarizes the electrons in the C-H sigma bond, leaving spin density
on the hydrogen that is opposite to the unpaired spin density in the carbon p.
orbital. The stabilizing feature of structure I over structure II is the electron
exchange interaction. It is the same effect that causes the lowest-energy excited
state for helium, 1s 12s', to be a triplet instead of a singlet. If we label the two
electrons that are mainly on carbon in each of the structures in Fig. 9-15,
we see that they can be interchanged in structure I, without changing the m,
value of the unpaired electron:
al b1
C1 H C1 H
b a
This interchange is not possible in structure II. Thus, I is stabilized by a quantum

mechanical interaction analogous to resonance.
A more complete molecular orbital description of this effect is presented in
Dewar's text!') In treating the e 2 r;j interactions in a molecule, one type of integral
that is obtained describes a repulsive interaction and has the form
J.,= f,*(jir(i) - 0,*(j)VI(j) dzi dzc (9-9)
This represents the Coulomb repulsion of the electron density [j f,*(i)O,(i)dTj

is the density in i] from electrons i and j in orbitals m and n, where m may or
may not equal n. There are other integrals that we shall label Kmn, which are
zero when the electron spins are paired and are non-zero when the spins are
parallel. The K., integrals are:
K., j m*(j),(m) e
ij
f,*(j)f,(j)dt i dT, (9-10)
This is called an exchange integral because it corresponds to an exchange of the

orbitals containing electrons i and j. When the spins of the two electrons are
parallel, the squares of these wave functions show that the two electrons have a
drastically reduced probability of being near each other when compared to two
electrons with opposite spins. Thus, the Coulomb repulsion is decreased when
the spins are parallel because the electrons stay away from each other. The
magnitude of the K., integral depends upon the overlap of the two orbitals m
and n. The quantity ,(j)O(j) dr1 was referred to earlier as the differential overlap.
The overlap integral of the differential overlap of two orthogonal orbitals is zero;
i.e.,
fO(j)O,*(j)dT= 0
However, when the differential overlap in the volume element dTr is operated
upon by e2 , multiplied by O,*(i)On(i), and integrated over all volume elements
dz, and dr1 , the result is not zero, but is the exchange integral. These exchange
effects not only give rise to the spin polarization described above and cause the
first excited triplet state of helium to be lower than the first excited singlet, but
they also lead to Hund's rule.
The spin density experienced at the hydrogen atom of the C-H bond when
there is unpaired electron density in the it-orbital (2p.) is expressed by equation
(9- 11):
aH Q (9-11)
where pc is the unpaired electron densit' in the carbon 2 p, orbital and Q is the
value of aH when there is a full electron on the carbon. The spin density at the
proton is negative, so aH is negative and Q must be negative. In systems where
Pc is known, Q can be calculated and is found experimentally to vary from -22
to -27 gauss. A rough value of -23 gauss for aromatic radicals treated by
Hickel theory suffices for most purposes that will concern us. When an extended
Hilckel calculation is used, the overlap is not set equal to zero and the molecular
orbital coefficients are normalized to include overlap. Accordingly, the value of
Q used depends on the m.o. calculation used, that is, on the definition of pc
When there is a node in the m.o. containing the unpaired electron at one
of the carbon atoms in a it-system, similar exchange interactions with lower-energy
filled pi-molecular orbitals operate to place negative spin density on this carbon.
We shall make this more specific in the discussion of the allyl radical below.
Resulting exchange interaction of this unpaired spin with the C-H sigma bond
places positive spin density on the hydrogen. The Hickel, extended HUckel, or
any restricted m.o. calculation (i.e., one where two electrons are fed into each
molecular orbital) do not include these exchange interactions. They simply
indicate a node at the carbon atom or the hydrogen atom. One attempts to
correct for this shortcoming, for example, at a hydrogen directly bonded to a
carbon containing unpaired electron density in an orthogonal. C,, orbital by
employing equation (9-11). Often other polarization effects in a molecule are H (4.06)
qualitatively discussed, and the protons or carbons where these effects dominate 2
(13.93) H C H (13.93)
are ignored when one attempts a quantitative fit of the calculated and experimental
coupling constants in a molecule. This discussion can be made more specific by (14.98) H H (14.98)
considering the observed"' proton hyperfine splittings in the allyl radical shown
in Fig. 9-16. The radical contains three electrons in the it-system whose wave FIGURE 9-16 Proton
functions are given by: hyperfine splittings in the
allyl radical.
01 = 2 ((Pi + 2/292 + 93) (bonding) (9-12)
r2 = 2(1i P3) (non-bonding) (9-13)

1
3 ((PI - 292 + 93) (anti-bonding) (9-14)
The odd electron is placed in 02, so one would predict the unpaired density at
C1 to be pc, = (1//2)2 = 0.5, where 1/2 is the C, coefficient in the m.o.
containing the unpaired electron. Using equation (9-11) with Q = -23, one would
predict a, to be equal to -11.5. Furthermore, there would be no unpaired
electron density at C2 and, without some kind of spin polarization involving the
carbon n electron density, one would predict a zero coupling constant for this
hydrogen or for a 3 C at this position. This is not observed. Two polarizations
are needed to account for a hydrogen coupling constant from this middle
hydrogen. The qualitative explanation involves taking the filled orbital 0, and
writing two separate spin orbitals for it, 0/, and 0/a,. Only one electron is placed
in each spin orbital. The wave functions in terms of the foregoing wave functions
for allyl become
01ia - V/1 + i{V 3
/l -- 01 - AV 3 (where A << 1)
For spin aligned with the field in the lower energy V1, and opposed to the field
in 01b, there will be increased spin density aligned with the field on atoms 1 and
3 relative to that on atom 2. In 0 1 , with spin density opposed to the field, there
will be more negative spin density on atom 2 than on carbon atoms 1 and 3. By
this mechanism, we are not introducing any unpaired electrons into the old 0,
orbital, but we simply are influencing the distribution of the paired spins over
the three atoms giving rise to negative (opposed to the applied field) spin density
on C2. This negative spin density then undergoes spin polarization with the
electron pair in the C-H bond [see the discussion of equation (9-11)] to place
spin density on the hydrogen. The exchange interaction of the unpaired electron
in 0 2 (mainly on C1 and C3) with the pair in V1 is the effect that lowers the
energy of V1, relative to V/1b. The two hydrogens on one of the terminal carbons
are not equivalent by symmetry, but our discussion so far has not introduced
any effects that would make them non-equivalent from the standpoint of spin
distribution. Exchange polarization involving filled sigma molecular orbitals is
required.
The phenomena discussed above are all indirect mechanisms for placing
unpaired spin density on the hydrogen. When the free radical is a sigma radical,
e.g., the vinyl radical H2C=C-H, the protons in the molecule make a
contribution to the sigma molecular orbital containing the unpaired electron.
Thus, the unpaired electron is delocalized directly onto the proton and aH is
proportional to 02. Since Hckel calculations are inappropriate for sigma systems
of this sort, the initial work in this area utilized extended Huckel molecular
orbital calculations. Procedures have been reportedm') for evaluating 0 2 at the
nucleus from the wave function, and aH is calculated by:
2
aH (gauss) = 1887 O(H) (9-15)
Again, spin polarization is foreign to this calculation. Often the majority of the
unpaired electron density resides at a given atom in the radical. When this is the
case, spin polarization makes a large contribution at protons directly bound to
this atom. Poor agreement between calculated and experimental results can be
expected for this atom. Generally, for other protons in the molecule (with a few
exceptions), spin polarization effects make a relatively insignificant contribution
when direct delocalization is appreciable.
Semi-empirical quantitative approaches(8 91 ) have been reported to incorpor-
ate the effects of spin-polarization. An unrestricted molecular orbital calculation,
i.e., one utilizing spin orbitals, yields the best results. The most common one at
present is the so-called INDO calculation,(9) which has been parametrized to
calculate spin densities on hydrogen. It has not been extensively tested on atoms
for which spin polarization dominates, but is certainly the method of choice for
this calculation at present. The output"O) consists of one-electron orbitals, and
all the positive and negative spin densities at an atom in all the filled molecular
orbitals are summed to produce the net spin density at the atom.
The application of the results from molecular orbital calculations to the
hyperfine splittings from atoms other than hydrogen is considerably more
complex. In contrast to protons, which have only the direct and indirect
mechanisms described above, 13 C hyperfine splittings have contributions from
other sources. (1) Unpaired electrons in a p(ni) orbital can polarize the filled 2s
and the filled Is orbitals on the same atom. (2) There can be direct delocalization
of electron density into the 2s orbital in a sigma radical. (3) Spin density on a
neighbor carbon, by polarizing the C-C sigma bond, can place spin density into
the 2s and 2p orbitals of the carbon whose resonance is being interpreted. The
calculations(1 -13) of ' 4N, 13S, and 170 hyperfine coupling have been more
successful than those for 1 3 C. Silicon-containing radicals have also been success-
fully treated.!1 3) The effects from spin densities on neighboring atoms are found
to be less important for these nuclei than for "C.
The application of the results from molecular orbital calculations to the
assignment of the esr spectrum of an organic radical is an important application
of the foregoing discussion. Another application involves determining the
geometry of free radicals. For instance, is CH 3- planar? Does the C-H bond of
the vinyl radical lie along the C-C bond axis? When the calculated hyperfine
coupling constants are found to vary considerably with geometry (i.e., a whole
series of molecular orbital calculations are performed for different geometries),
the fit of calculated and experimental results can be used to suggest the actual
geometry. (14) In several examples, molecular orbital calculations have provided
evidence about the structure of a radical produced in an experiment. Ila, 14) For
example, y irradiation of pyridine produced a radical believed to be the pyridine
cation; i.e., one of the lone pair electrons was removed. The results shown in
Table 9-2 indicate that the 2-pyridyl radical was actually formed." ia)
TABLE 9-2. Extended HOckel and Experimental Isotropic Hyperfine Coupling

Constants (Gauss)
Calculated for
Calculated for 2-pyridyl
C5HN+ CH 4 N Experimental
aN 52.5 33.8 29.7

aH 27.0 (2,6 H) 5.3
9.7 (3.5 H) 9.3
38.8 (4 H) 7.6 4.3
0.0
These calculations and esr experiments indicate the very extensive amount
of electron delocalization that occurs in the sigma system. For example, significant
amounts of unpaired electron density are found on the methyl group protons of
CH 3NH 2 + and C 2 H 5NH 2 +. Thus, the lone pair molecular orbital is not a
localized nitrogen lone pair orbital, but is a delocalized molecular orbital. The
16
extent of delocalization onto the protons varies in the different rotamers.( )
Anisotropic 9-6 ANISOTROPY INTHE g VALUE

Effects The next feature of esr spectroscopy can be introduced by describing the g values
for the NO 2 radical trapped 1 7 ) in a single crystal of KNO 3. When the crystal
is mounted with the field parallel to the z-axis of NO 2 (the twofold rotation axis),
a g value of 2.006 is obtained. When the crystal is mounted with the x- or y-axis
(the plane of the molecule containing y) parallel to the field, a g value of 1.996
is obtained. The molecule is rapidly rotating about the z-axis in the solid, so the
same result is obtained for x or y parallel to the field. The differences in the g
values with orientation are even more pronounced in transition metal ion
complexes (vide infra) and in complexes of the lanthanides and actinides.
The treatment so far has involved so-called isotropic spectra. These are
obtained when the radical under consideration has spherical or cubic symmetry.
For radicals with lower symmetry, anisotropic effects are manifested in the solid
spectra for both the g values and the a values. Usually, for these lower symmetry
systems, the solution spectra appear as isotropic spectra because the anisotropic
effects are averaged to zero by the rapid rotation of the molecules. Our concern
here is how these anisotropic effects arise and how they can be determined. Later
(Chapter 13), we shall see how the anisotropy in g and a can be used to provide
information about the electronic ground state of transition metal ion complexes.
Anisotropy in g arisesfrom coupling of the spin angular momentum with the
orbitalangularmomentum. The spin angular momentum is oriented with the field,
but the orbital angular momentum, which is associated with electrons moving
in molecular orbitals, is locked to the molecular wave function. Consider a case
where there is an orbital contribution to the moment from an electron in a
circular molecular orbit that can precess about the z-axis of the molecule. In Fig.
9-17, two different orientations of this molecular orbital relative to the field are
indicated by ellipses. In Fig. 9-17(A), ItL (the orbital magnetic moment vector*)
and I, (the spin magnetic moment vector) are in the same direction. In Fig.
9-17(B), a different orientation of the molecule is shown. The electron moment
ft, has the same magnitude as before, but now the net moment, indicated by the
boldface arrow, results because ft, and 4, do not point in the same direction.
If it were not for the orbital contribution, the moment from the electron would
be isotropic. When the effects of the orbital moment are small, they are incorporated
into the g-value; and this g-value will be anisotropic.A tensor, equation (9-16), is
needed to describe it. The g-tensor than gives us an effective spin, i.e., g - S = S,,,.
For different orientations, the g-tensor lengthens and shortens Se, to incorporate
orbital effects. It should be emphasized that even when the ground state of a
molecule has no orbital angular momentum associated with it, field-induced
mixing in of an excited state that does have orbital angular momentum can lead
to anisotropy in g. The.information that g-tensor anisotropy provides about the
electronic structure of the molecule will be discussed in Chapter 13 on the epr
of transition metal ion complexes.
*The direction of the orbital angular momentum given by the right-hand rule is opposite to
that of the magnetic moment vector of the electron.
9-6 Anisotropy in the g Value 381
As FIGURE 9-17 Coupling of

Ss the projections of the spin
and orbital angular
IYL momentum for two different
e molecular orientations
H relative to the applied field H.
(A) (B)
For an isotropic system, we wrote our Hamiltonian to describe the interaction

of an electron spin moment with the magnetic field and with a magnetic nucleus
in equation (9-4) as:
H g#3HS- gNflNHI + ab S
Both g and a were scalar quantities. When an anisotropic free radical is

investigated in the solid state, both g and a have to be replaced by tensors or
matrices. The g#3HS, term in the Hamiltonian becomes #-g -H, which can be
represented with matrices as:
[gx g g,, LHx9

#(~S S, SZ] g,, g,, g, H, 9-6
gz, gz, g,,_ Hz_
and a in the aI -S term is also replaced by a tensor. Here x, y and z are defined
in the laboratory frame; i.e., they are crystal axes. The off-diagonal element g,
gives the contribution to g along the z-axis of the crystal when the field is applied
along the x-axis. This matrix is diagonal when the crystal axes are coincident with
the molecular coordinate system that diagonalizes g. When they are not coincident
and the crystal is studied along the x, y, and z crystal axes, we get off-diagonal
contributions, as we shall show subsequently. The g matrix can be made diagonal
by a suitable choice of coordinates.
If one studies the esr of a single crystal with anisotropy in g, the measured
g value is a function of the orientation of the crystal with the field because we
measure an effective g value oriented along the field. If we define molecular axes
X, Y, and Z that diagonalize the g-tensor and pick as an example a case in which
they are coincident with the crystal axes, the effective values of g for an arbitrary
orientation of the crystal is then given by
g = (2 2 ) cos 2 H + (g2 ),, cos 2 0

Hy + (g2 ) cos 2 H,
2 2 2 2 2 2
= (g )jl + (g )gl Y + (g ) jl (9 17)
Here 6Xn, 6,, and 0,, are the angles between the field H and the X-, Y-, and Z-
axes, respectively. The symbols l, l,, and l are often used to represent the cosines
of these angles and are referred to as the direction cosines. From trigonometry,
k2 + /,2+ l.2 = 1, so two parameters suffice to specify the direction. The above
equation can be indicated in matrix notation as
(gI2), o 0
genf2 = l ,l]0 (g 2),, l, (9-18)
Y j 0 0 (g2 I
Since S,, S,, and S. as well as H., H,, and H. are defined in terms of the molecular
coordinate x,y,z system, they can be replaced by the same direction cosines. The
molecular coordinate system that diagonalizes the g-tensor may not be coincident
with the arbitrary axes associated with the crystal morphology. Since this
experiment is carried out using the easily observed axes of the bulk crystal, the
above equation has to be rewritten in non-diagonal form as
[Wg). Wg)k, (q)2

(gq
1[X]
)yz I I (9-19)
q ff 2 = l i ,2 (
(g2) -
,q )Y (g2
Equation (9-19) can be used to evaluate all the tensor components. The matrix
is symmetric [i.e., (g2 )= (g2 ),,], so only six independent components need to
be evaluated. Since it is most convenient to orient the crystal in the magnetic
field relative to the observed crystal axes, the x, y, and z axes are defined in terms
of these observed axes of the bulk crystal. Sx, S,, and Sz, as well as H,, H,, and
Hz, are defined in terms of these axes. Consider the first case where the crystal
is mounted, as shown in Fig. 9-18, with the y-axis perpendicular to the field so
that the crystal can be rotated around y with H making different angles, 0, to z
in the xz-plane. Now, 1z equals cos 0 and I, equals sin 0, where 0 is the angle
between H and the z-axis. Substituting these quantities into equation (9-19) and
carrying out the matrix multiplication yields
2(g 2) sin 0 cos 0 + (g2 )zz cos 2

geff2 = ( 2
),, sin 20 + 0 (9-20)
FIGURE 9-18 Mounting of a

crystal for rotation in the xz-plane. H
For rotations in the yz-plane, we have l = 0, l, = sin 0, and 1z = cos 0. Substitu-
tion into (9-19) and matrix multiplication then yields
gff = (g 2 )Yy sin 2 0+ 2(g 2 ),2 sin O cos 0+ (g 2 )zz cos 2 0 (9-21)
In a similar fashion, rotation in the xy-plane yields
g,2 2
(g ) , cos 2 0 +2(g2),, sin Ocos 0+ (g 2 ),, sin 2 0 (9-22)
These equations thus tie our matrix in equation (9-19) to experimental

observables, gff2. In our experiment, a g-value is obtained, but we do not know
its sign. The measured g-value is squared and used in this analysis. For rotations
in any one plane, only three measurements of ge,, 2 need be made (corresponding
to three different 0-values) to solve for the three components of the g 2 -tensor in
the respective equations. For the xz-plane, one measures (g2 ),, at 0 = 0 and (g2)
at 0= 90'. With these values and ge, 2 at 0= 45', one can solve for (g2 ) 2. In this
way, the six independent components of the g 2 -tensor can be measured. In practice,
many measurements are made and the data are analyzed by the least squares
method. One then solves for a transformation matrix that rotates the coordinate
system and diagonalizes the g 2 -tensor. This produces the molecular coordinate
system for diagonalizing the g 2 -tensor, and the square roots of the individual
diagonal g2 matrix elements produce g.., g and g.. in this special coordinate
system. In order for this procedure to work as described, it is necessary that all
the molecules in the unit cell have the same orientation of their molecular axes
relative to the crystal axes. Thus, these measurements are often carried out in
conjunction with a single crystal x-ray determination.
9-7 ANISOTROPY IN THE HYPERFINE COUPLING
We introduced Section 6 by describing the anisotropy in g when a single crystal

of NO 2 in KNO 3 was examined at different orientations relative to the field. The
a-values of this system are also very anisotropic. When the molecular twofold
axis is parallel to the applied field, the observed nitrogen hyperfine coupling
constant is 176 MHz, while a value of 139 MHz is observed for the orientation
in which this axis is perpendicular to the field. In rigid systems, interactions
between the electron and nuclear dipoles give rise to anisotropic components in
the electron nuclear hyperfine interaction. The classical expression for the
interaction of two dipoles was treated in Chapter 8, and the same basic
considerations apply here. For the interaction of an electron moment and a
nuclear moment, the Hamiltonian is:
5 7 3(5 -i)(1-4)
Hdipolar = -- g _r3 3Nj(N( (9 23)
The sign is opposite that employed for the interaction of two nuclear dipoles,
which was the problem treated in the section on liquid crystal and solid nmr.
Substituting 5 = 5 + , + Sz, = + Z, + 1z, and r = x + y + z, and expand-

ing these vectors, leads to
- 3X2 +- I r 2 5--3y2 - + [r 2 r53]Z2

Hdipolar = gN I{[r2
N3a,- + gNN) x x S ) 5 z 3y)

z+
(9-24)
When this Hamiltonian is applied to an electron in an orbital, the quantities in
brackets must be replaced by average values; we employ angular brackets to
refer to the average value over the electronic wave function. In matrix notation,
we then have
HdipoIar
r2 -3X2 -3xy 3xz

r 2s rK r
-3xy) r2 _ 3y2 -3yz Iy (9-25)
-(g~gN#N[x 3y z r r 5r5 (-5
-3xz 3yz r2 - 3z2
This equation is abbreviated as
Hdipoiar =h - T (9-26)
where T is the dipolar interaction tensor (in units of Hz) that gauges the anisotropic
nuclear hyperfine interaction. The Hamiltonian now becomes
R = #3 -g -H -gNNH -Z + hS-A -Z (9-27)
where the first term on the right is the electron Zeeman term, the second is the
nuclear Zeeman term, and the third is the hyperfine interaction term. The quantity
A in the third term includes both the isotropic and the anisotropic components
of the hyperfine interaction; i.e.,
A= T + a1 (9-28)
In the application of the Hamiltonian given in equation (9-27) to organic free

radicals, several simplifying assumptions can be introduced. First, gN/NH -I, the
nuclear Zeeman effect, usually gives rise to a small energy term compared to the
others. (Recall our earlier discussion about the energies of the esr and nmr
transitions). Second, g-anisotropy is small, and we shall assume that g is isotropic
in treating the hyperfine interaction.* (This would be a particularly bad
* If g-tensor anisotropy is comparable to hyperfine anisotropy, this assumption cannot be made.

The reader is referred to Chapter 13 and to A. Abragam and B. Bleany, "EPR of Transition Ions,"
p. 167, Clarendon Press, Oxford , England, 1970 for a discussion of this situation.
assumption for certain transition metal complexes, vide infra.) The electron
Zeeman term is assumed to be the dominant energy term, so 5 is quantized along
H, which we label as the z-axis. We see in this example, as we shall see over and H
over again, that it is often convenient to define the coordinatesystem to be consistent
with the largest energy ,ffect. Next we have to worry about the orientation of
the nuclear moment relative to the z-field. Our discussion is general, but it may
Hb
help to consider an ethyl radical oriented as shown in Fig. 9-19, with the CH H/
3
group not undergoing rotation. To make this point, focus attention on nucleus
H. involved in dipolar coupling to the electron. The nuclear moment will not be
quantized along z, but along an effective field, Herr, which is the vector sum of
the direct external field H (nuclear Zeeman) and the hyperfine field produced by FIGURE 9-19 The
the nearby electron. If the hyperfine interaction is large (- 100 gauss), the hyperfine orientation of the spin and
nuclear moments in an
field at this nucleus (i.e., the field from the electron magnetic moment felt at the
applied field.
hydrogen nucleus) is about 11,700 gauss. (This is to be contrasted to the field of
~3000 gauss from the magnet and the field of ~18 gauss felt at the electron
from the considerably smaller nuclear moment.) Thus, we may be somewhat
justified in ignoring the nuclear Zeeman term, g 4NH - Z, in equation (9-27). The
Hamiltonian for most organic free radicals (where g is isotropic) is then
considerably simplified from the form in equation (9-27), and becomes
H = g/)HS + hSz(AzZ + AnZJ + AZ) (9-29)
T he terms on the far right give the z-component of the electron-nuclear hyperfine
interaction with contributions from Z. and Zy as well as from Z, for the z-field
does not quantize I, but does quantize S. When this Hamiltonian operates on
the Iz)c and other wave functions, off-diagonal matrix elements in the secular
determinant result. When it is diagonalized and solved for energy, the following
results are obtained:
1
g/3H + I1 /_A 2
+ A,, 2 +A 2
YN} E 2 4
#N } I
2
I
g/3H +-z2+ A
24
z2+A2
.
The term containing the square root replaces the ('%,)a obtained from the
evaluation of the af -$ term for the hydrogen atom. The energy for the hyperfine
coupling is thus given by
AEhf 2 A + A_2
y ± A-2 (9-30)
The quantity A contains both the isotropic, a, and anisotropic, T, components

of the hyperfine interaction. Since in solution the anisotropic components are
averaged to zero, it becomes a simple matter to take one-third of the trace of A
to decompose A into T and a. (This assumes that the solvent or solid lattice has
no effect on the electronic structure.)
These expressions apply for any orientation of the molecule relative to the
applied field. In a single-crystal experiment, in which the crystal and molecular
axes are not aligned, we proceed as in the case of the evaluation of the g-tensor
to determine all of the components of the hyperfine tensor. The coordinate system
that diagonalizes the g-tensor need not be the same one that diagonalizes the
A-tensor, and neither one of these need be the apparent molecular coordinate
system."") If the molecule has overall symmetry (i.e., the full ligand environment
included) such that it possesses an n-fold rotation axis, the same axis will be
diagonal for g and A, and it must be coincident with the molecular z-axis.
The angular dependence of the hyperfine interaction for the case where the
field from the hyperfine interaction is large, I = 1/2, and the system has axial
symmetry can be expressed by substituting r cos 0 for z and r sin 0 for x and y
into equation (9-24). We are, in effect, resolving the nuclear moment in Fig. 9-19
into components parallel and perpendicular to the field. The Hamiltonian
including the electron Zeeman term (g#H z) becomes
R = g#Hgz + h9z{[a + B(3 cos 2

0 - 1)]fz + 3B cos 0 sin OI} (9-31)
The result of this Hamiltonian operating on the basis set produces the energies
given by
hM
E = gpHM, h S [(a - B)2 + 3B(2a + B) cos2 0] 2 (9-32)
2
where a is the isotropic hyperfine coupling constant, B is the anisotropic hyperfine

coupling constant, and 0is the angle that the z-axis of the molecule makes with
the field. The hyperfine coupling constant A observed experimentally is the
difference between the energies of the appropriate levels and is given (in cm 1)by
A = h[(a - B)2 + 3B(2a + B) cos 2 01 2
One often sees the following equation presented in the literature to describe
the anisotropy of g and A for an axial system:
AE = hv = g + ggi)#Ho + am, + (g 1 - gi)#Ho + Bmi (3 cos 2 0 - 1)

(9-33)
where 0 is the angle between the z-axis and the magnetic field, a is the isotropic
coupling constant, and B is the anisotropic coupling constant. The equation
results from equation (9-32) by adding the anisotropy in g and by assuming that
both the g- and A-tensors are diagonal in the same axis system.
Since in the analysis of the anisotropy in the hyperfine coupling we deal
with A 2 [see equation (9-30)], usually one cannot obtain the sign of the coupling
constant from the esr experiment.* However, it can be readily predicted for an
organic radical that contains an electron in a p-orbital (there is no anisotropic
contribution from unpaired electron density in a spherical 2s orbital). We shall
begin by returning to equation (9-25) to further explore its meaning. Assume
* For certain systems, one can determine the sign of the coupling constant by using nmr, because
spin aligned with the field causes a downfield shift and spin opposed to the field causes an upfield shift.
that the electron is in a hypothetical orbital that can be represented by a unit
vector. When this hypothetical orbital lies along z, we have z = r, x = 0, and
y = 0, and all off-diagonal terms are zero. We observe, on substitution into
equation (9-25), that T,, = k<1/r 3 >, T, = k<1/r3 >, and T = k(2/r 3>.t [The
matrix elements as written in equation (9-25) have the opposite sign, but the
whole term is negative to describe the interaction of the positive nuclear moment
and the negative electron moment.] Note that the trace is zero.
To make the problem more realistic, we shall next consider the electron to
be in a p. orbital. It is convenient to convert to spherical polar coordinates to
solve this problem by substituting z = r cos 0, x = r sin 0 cos p, and y = r sin 0
sin cp. The result for an electron at a specific (r, 0) (after consideration of the
negative sign) is
3 -
Tz = g9#gNN<( c2 0 3
1
T,, 2I gN N(( 3
COS 2 0 - 1)/r 3 >
1
Tx = 2I9#gN#N( 3
2
cOS
2
0 - 1) r3 )
The latter two matrix elements are readily obtained by substituting for x and y
in equation (9-25) and substituting <cos 2 0> '/2 for an axial system. Note that
the trace is zero. Now, consider that the electron can be located at any place in
the p orbital. Thus, we have to integrate over all possible angles for the radius
vector to the electron in this orbital and then over all radii r. In doing so, we get
T, = g #g <r> 3(9-34)
3
where (1/r > is the average value of the quantity 1/r 3 . Abbreviating equation
(9-34) as T = 4 /P,, we have
2 2
T = 5 P, and T,, - P (9-35)
5 5
These considerations are valuable in predicting the signs of the anisotropic

components of the hyperfine coupling constant.
It is informative to apply these equations to the anisotropic hyperfine tensor
of the 3 C nucleus, which depends mainly on the unpaired electron density in
the p orbital of this atom. We wish to consider the signs of T., T,, and T, for
this system. The three orientations of the p-orbital in the molecule relative to the
applied field are indicated in Fig. 9-20. The dotted lines indicate the regions
where the (3 cos 2 0 - 1) function is zero. This correspondsto plotting the signs
for the various regions of the lines of flux emanating from the nuclear moment.
Accordingly, by visual inspection, we can tell whether T., expressed in equation
(9-34) will be positive or negative. For example, as we see in Fig. 9-20(A), when
f The proportionality constant is g9Ng/fh.

388 Chapter 9 Electron Paramagnetic Resonance SpectroscopY
--------------------------
----------------------------------------------- --------------------------
---------------------------------------------------------------------------
't ,--------------------------------------------
------------------- I-------------------
(A) (B) (C)
FIGURE 9-20 Visual representation of dipolar averaging of the electron and

nuclear moments: (A) p orbital oriented along the field; (B) and (C) p orbital
perpendicular to the field.
the p. orbital is aligned with the field almost the entire averaging of the dipolar
interaction of the nuclear moment over the p. orbital will occur in the positive
part of the cone. A large, positive T,,-value is thus expected. For the orientation
along the x-axis shown in Fig. 9-20(B), the dipolar interaction, T.,, will be large
and negative; the same is true of T, for the orientation shown in Fig. 9-20(C).
Analysis(18,19) of the "C hyperfine structure of the isotopically enriched malonic
acid radical, H'"C(COOH) 2

, produces a, = 92.6 MHz, Tx = - 50.4 MHz,
and T, = + 120.1 MHz. After the hyperfine tensor is

T, = -59.8 MHz,
diagonalized, the relative signs of T.., T,,, and T., are known (the trace must
be zero), but the absolute signs are not; i.e., all those given above could be
reversed. However, the arguments based on Fig. 9-20 provide us with good reason
to think that the signs presented above are correct.

It is informative to predict the signs of the anisotropic hydrogen hyperfine
components of a C-H radical. By analogy to our discussion above, the three
orientations of the p.-orbital of this radical shown in Fig. 9-21 predict that T,
is small, whereas T, is positive and T, is negative. Visual averaging of the
p orbital with our cone of magnetic nuclear flux also suggests that T, will be
small. Note that the cones representing the nuclear moment lines of flux are
drawn at the nucleus whose moment is causing the splitting via the dipolar
+ - -+ S+
------------------- ---------------- --- -------------------- --
fc I HI"I,-H
H, - H
H+
S + A+
--------------------------------------------------------------------------- ------------------------
(A) (B) (C)
FIGURE 9-21 Visual representation of the dipolar averaging of the nuclear

moment on the hydrogen with the electron moment in a p-orbital of carbon. (A) He
is parallel to the z-crystal axis; (B) He is parallel to the y-crystal axis; (C) H is
parallel to the x-crystal axis, and we are looking down the axis of the p-orbital.
interaction with the electron. If the x, y, and z axes are defined relative to the
fixed crystal axes (which are coincident with the molecular axes) as in Fig. 9-21,
calculation 2( ) shows that a full unpaired electron in the carbon p-orbital would
lead to an anisotropic hyperfine tensor of
-38 0 O
TH= 0 +43 0 MHz
The experimental proton hyperfine tensor for the a proton of the malonic acid
radical was found to be
AH = 0 +29 0 MHz
00+581
Since the isotropic hyperfine coupling constant, a, is one-third the trace of AH,
it equals + 59 MHz. Accordingly, the anisotropic hyperfine tensor T must be:
-32 0 01 32 0 01
TH 0 +30 0 MHz or 0 -30 0 MHz
0 0 +1 0 0 1
The arguments presented in the discussion of Fig. 9-21 and comparison to the
theoretical tensor lead us to predict the tensor on the left to be correct. This
matrix arose from a = - 59 MHz. Since a positive a would have given tensor
components that correspond to values greater than that for a full electron, the
isotropic hyperfine coupling constant must be negative.
Anisotropic and isotropic hyperfine coupling constants have been measured
in several organic and inorganic radicals and have provided considerable
information about the molecular orbital containing the unpaired electron. The
value of B for one electron in a p orbital of various atoms can be evaluated using
an SCF wave function from
B h gfgNfN{r-' (9-36)
5
For "C, the anisotropic hyperfine coupling constants are given by
+2B
where B is calculated from SCF wave functions to be 91 MHz.'For HC(COOH) 2 ,
the experimental value of T., for "C is found to be + 120.1 MHz compared to
the value of 182 MHz for an electron localized in a C2 p orbital. Accordingly, it
is concluded that p, is 0.66. From "C enrichment, it is found that a, = +92.6
MHz. A full electron in an s-orbital has an isotropic hyperfine coupling constant
of 3110 MHz. The measured a, corresponds to a value of 0.03 for C 2 s spin density.
The radical is expected to be nearly planar.
The magnitude of the isotropic "C hyperfine coupling constant supports a

planar("' CH 3 radical, a, = 38.5 gauss, but the value of a, = 271.6 gauss in the
CF 3 radical indicates(") that it is pyramidal with s-character in the orbital
containing the unpaired electron.
The isotropic "N hyperfine coupling constant in NO 2 is 151 MHz, and the
maximum value for the anisotropic hyperfine coupling constant is 12 MHz. With
1540 MHz expected for one electron in a nitrogen 2s orbital and 48 MHz for
an electron in a 2p orbital, p, is calculated to be 0.10 and p, is found to be 0.25,
for a 2p/ 2 s ratio of 2.5. An sp 2 orbital would have a ratio of 2.0, so this suggests
that more p-character is being used in the orbitals to bond oxygen and an angle
greater than 120 is predicted. Microwave results gave a value of 134 for NO 2
in the gas phase.
9-8 THE EPR OF TRIPLET STATES
The next complication we shall discuss arises when there is more than one
unpaired electron in the molecule. An example is provided by the triplet state
that is formed upon u.v. irradiation of naphthalene. The single crystal epr spectrum
was studied for a sample doped into durene. The similar shapes of these two
molecules allowed the naphthalene to be trapped in the durene lattice; dilution
of the naphthalene greatly increases the lifetime of the triplet state. The spectrum
consists of three peaks, which changes resonance fields drastically with orientation
of the crystal. The changes could not be fitted with the anisotropic g- and a-tensor
we have developed. The anisotropy in this system arises from electron-electron
spin interaction and is described by the spin Hamiltonian given in equation
(9-37). This Hamiltonian is seen to be very similar to that for the dipolar
interaction of an electron and nuclear spin [equation (9-23)].
#H + ,) + 2#
FI =- 3, {$1 - $2 3($, _- 32)(52'
2
1 + $2 ) + g /
g3H($ 2
3( s (9-37)
where r is the vector joining the two electrons labeled 1 and 2.
FIGURE 9-22 (A) Singlet ground

term; (B) ms = 1, 0, -1
components of the S = 1 state.
m 8 =0 mn=1 m0=0 m= -1
(A) (B)
9-8 The EPR of Triplet States 391
+1+
+1
Ms +1, 0, -1
D
-1 00
H increasing -- H increasing H increasing -

0 0 0
(A) (B) (c
FIGURE 9-23 The effects of zero-field splitting on the expected epr transitions. (A)
No zero-field effects. (B) Moderate zero-field splittings. The dashed arrows show
the fixed frequency result, and the solid arrows show the fixed field result. (C)
Large zero-field effects. The magnetic field is assumed to be parallel to the dipolar
axis in the molecule.
The magnitude of the contribution to the epr spectrum from these effects
depends upon the extent of the interaction of the two spins. We discussed the
Coulomb (J) and exchange (K) integrals earlier in this chapter. When two
molecular orbitals are closer in energy than the difference between exchange and
repulsive energies, a stable triplet state arises. In the case of naphthalene, the
ground state is a singlet, and the excited triplet state has an appreciable lifetime
because of the forbidden nature of the triplet-singlet transition. For the ground
state S = 0 and ms = 0, and this configuration is illustrated in Fig. 9-22(A). For
the triplet state we have S = 1 and m, = 1, 0, -1. These electronic configurations
are illustrated in Fig. 9-22(B). If orily exchange and electrostatic interactions
existed in the molecule, the three configurations ms = 1, 0, and -1 would be
degenerate in the absence of a magnetic field. The magnetic field would remove
this degeneracy as shown in Fig. 9-23(A), and only a single transition would be
observed as was the case for S = 2. However, magnetic dipole-dipole interaction
between the two unpaired electrons removes the degeneracy of the ms components
of S = 1 even in the absence of an external field, as shown in Fig. 9-23(B). This
removal of the degeneracy in the absence of the field is called zero-field splitting.
When a magnetic field is applied, the levels are split so that two Ams = + 1
transitions can be detected, as illustrated in Fig. 9-23(B). Earlier we mentioned
that three peaks were observed for the epr spectrum of triplet naphthalene. Two
of these are the Ams = + 1 transitions shown in Fig. 9-23(B); the third is the
Ams = ± 2 (between ms = --1 and ms = + 1) transition, which becomes allowed
when the zero-field splitting is small compared to the microwave frequency. When
this is the case, one cannot assign precise ms = + 1, 0, or - 1 values to the states
that exist. When the zero-field splitting is very large, as in Fig. 9-23(C), the ms
values become valid quantum numbers and the energies for the Ams = ± 1 allowed
transitions become too large to be observed in the microwave region; accordingly,

no spectrum is seen.
Since the electron-electron interaction is dipolar [equation (9-37)], it is
expected to be described by a symmetric tensor, the so-called zero-field splitting
tensor, D.
HD = S -D-S (9-38)
The D-tensor elements have the same form as those for T given in equation (9-25).
This dipolar D-tensor accounts for the large anisotropy observed in the spectrum
of a bi-radical. In terms of the principal axes that diagonalize the zero-field
splitting tensor, we can write
NII = - XS|2 - Y - Z$|2
where X is the D. element in the diagonalized zero-field tensor. Since the tensor
is traceless, (X + Y + Z = 0), the zero-field Hamiltonian can be written in terms
of two independent constants, D and E:
S= D( $ 2 5-$ + E(- 2
_ 2
)
Since X = Y in a system of axial symmetry, the last term disappears. If the

molecule has cubic symmetry, no splitting from this zero-field effect will be
observed.
Operating with this Hamiltonian on the triplet state wave functions, one
can calculate the energies as a function of field and orientation. The results
indicate substantial anisotropy in the spectrum. The spectrum for the naphthalene
triplet state is described by g (isotropic) = 2.0030, D/hc = + 0.1012 cm-', and
E/hc = - 0.0141 cm '. The magnitudes of D and E are related to how strongly
the two spins interact. These quantities become negligible as the two electrons
become localized in parts of the molecule that are very far apart.
In liquids, the traceless tensor D averages to zero. The large fluctuating fields
arising from the large, rotating anisotropic spin-spin forces in molecules with
appreciable zero-field splitting cause effective relaxation. The lines in the spectra
are thus usually too broad to be detected. With some exceptions, the epr of triplet
states cannot be observed in solution unless the two spins are far apart (i.e., D
and E are small).
9-9 NUCLEAR QUADRUPOLE INTERACTION
A nucleus that has a nuclear spin quantum number I 1 also has an electric
moment, and the unpaired electron interacts with both the nuclear magnetic and
electric moments. The electric field gradient at the nucleus can interact with the
quadrupole moment as in nqr, and this interaction affects the electron spin energy
states via the nuclear-electronic magnetic coupling as a second-order perturbation.
The effect of quadrupole interaction is usually complicated because it is
9-9 Nuclear Quadrupole Interaction 393
accompanied by a much larger magnetic hyperfine interaction. The orientation

of the nuclear moment is quantized with respect to both the electric field gradient
and the magnetic field axis. When the magnetic field and the crystal axes are
parallel, the only quadrupole effect is a small displacement of all the energy levels
by a constant amount, which produces no change in the observed transitions.
However, when the two axes are not parallel, the effect is a competition between
the electric field and the magnetic field. This has two effects on the spacing of
the hyperfine lines: (1) a displacement of all energy levels by a constant amount
and (2) a change in the separation of the energy levels that causes the spacing
between adjacent epr lines to be greater at the ends of the spectrum than in the
middle.
This quadrupole effect can easily be distinguished from another second-order
effect that produces a gradual increase or decrease in the spacing from one end
of the spectrum to the other. The variation in the spacing from this other
second-order effect occurs when the magnetic field produced by the nucleus
becomes comparable in magnitude to the external field. In this case, the unequal
spacing can be eliminated by increasing the applied magnetic field.
A further effect of this competition between the quadrupolar electric field
and the magnetic field is the appearance of additional lines that are normally
forbidden by the selection rule Am, = 0. Both Am, = 1 and Am, +2
transitions are sometimes observed( 2 2 ) An analysis of the forbidden lines gives
the nuclear quadrupole coupling constant."" The approach involves a single-
crystal epr study of a compound doped into a diamagnetic host. A spectrum
containing these transitions is illustrated ( 4 ) in Fig. 9-24 for bis (2,4-pen-
tanedionato)copper(II) [63Cu(acac)2 ], doped into Pd(acac) 2. The forbidden tran-
sitions are marked in Fig. 9-24(A) with arrows and the other bands are the four
allowed transitions [I(63 Cu)= '/2]. The spacing and intensity of the forbidden
lines vary considerably with the angle 0. The variation in the spacing is shown
in Fig. 9-24(B). By matrix diagonalization, the spectra could be computer-fitted14 1
to a spin Hamiltonian:
H = P[g ,H S + g1 (H $, + H,$,)] + AS z + B($X1J + S,!,)
+ Q 1Z3 1(I + 1) -NPNH -
The term Q'[I 2_ (1/3)I(I + 1)] accounts for the quadrupole effects, and all other
symbols have been defined previously. A value of Q'/hc = (3.4 + 0.2) + 10~4 cm - I
fits all the data; and for I = 3/2, 4Q' is the quantity related to the field gradient
e 2Qq (see Chapter 14) at the copper nucleus, i.e., Q' = 3e2 qQ/41(2I + 1). In similar
studies of Cu(bzac) 2 doped into Pd(bzac)2 and of bis-dithiocarbamato copper(II)
[Cu(dtc)2] in diamagnetic Ni(dtc) 2, Q'-values of (3.3 + 0.21) + 10-4 cm -1 and
(0.7 ± 0.1) x 10~4 cm -1, respectively, were obtained. A Q'-value of about
16 x 10 - cm is expected for a hole in a d _,2 orbital of a free copper ion,
and covalency should lower the value found (Q' measures the electric field gradient
in the direction of the symmetry axis). The charge distribution in the sulfur
complex is very close to symmetrical and was interpreted to indicate a very
considerable amount of covalency in the copper-ligand bond.
10 gauss
FIGURE 9-24 (A) The

0-band spectrum of
Cu(acac)2 at 0 = 87'.
Forbidden lines are
indicated by arrows. (B)
Angular dependence of
the spacing between each (A)
pair of forbidden lines
(Am,= +1) in the
spectrum in (A). The
curves are calculated
values using
0'= 3.4 x 10-4 cm-', and
the symbols are
experimental values. The
letters /, m, and h
represent low-field,
medium-field, and ca
high-field pairs, C'
respectively. [Reprinted U, 1
with permission from H. cJ50

So and R. L. Belford, J.
Amer. Chem. Soc., 91,
2392 (1969). Copyright by
the American Chemical
Society.]
45 60
Angle (degrees)
9-10 LINE WIDTHS IN EPR
In this section, we shall discuss briefly a number of factors that influence the epr
line width. Many of these are similar to effects we have discussed in nmr.
Broadening due to spin-lattice relaxation results from the interaction of the
paramagnetic ions with the thermal vibrations of the lattice. The variation in
spin-lattice relaxation times in different systems is quite large. For some
compounds it is sufficiently long to allow the observation of spectra at room
temperature, while for others this is not possible. Since relaxation times usually
increase as the temperature decreases, many salts of the transition metals need
to be cooled to liquid N 2 , H 2 , or He temperatures before well-resolved spectra
are observed.
Spin-spin interaction results from the small magnetic fields that exist on
neighboring paramagnetic ions. As a result of these fields, the total field at the
ions is slightly altered and the energy levels are shifted. A distribution of energies
9-10 Line Widths in EPR 395
results, which produces broadening of the signal. Since this effect varies as
(l/r3 )(1 - 3 cos2O), where r is the distance between ions and 0 is the angle between
the field and the symmetry axis, this kind of broadening will show a marked
dependence upon the direction of the field. The effect can be reduced by increasing
the distance between paramagnetic ions by diluting the salt with an isomorphous
diamagnetic material; for example, small amounts of CuSO 4 can be doped into
a diamagnetic host ZnSO 4 crystal.
As in nmr, rapid chemical processes also influence the spectral line widths
and appearance. Resonances that are separate in the stopped exchange limit will
broaden as the process rate increases, and they will then coalesce to give a
weighted-average single resonance. With one-half the width at half height of an
organic free radical being typically - 0.1 gauss, significant line broadening occurs
for processes with first order rate constants of 5 x 107 sec- .
Electron spin exchange processes are very common in free radical systems,
and these effects drastically influence line width and spectral appearance. In
solution, this is generally a bimolecular process in which two radicals collide and
exchange electrons. The effects are similar to those observed in nmr and are
illustrated in Fig. 9-25 for solutions of varying concentrations of di-t-butyl
nitroxide, [(CH 3) 3C] 2 NO. As the concentration increases on going from (A) to
(C), the rate of bimolecular exchange increases and the resonances broaden. The
rate constant for this process has been evaluated in the solvent N,N-dimethyl-
formamide and was found to be 7 x 10' 1 mole-1 sec -, a value corresponding
to the diffusion controlled limit. As the solution whose spectrum is shown in Fig.
9-25(C) becomes considerably more concentrated, the single line resonance
sharpens as in the very fast exchange limit in nmr. This sharpening is referred
to as exchange narrowing.
Electron transfer between a radical and a diamagnetic species can also occur
at a rate that causes line broadening of the epr spectrum. One of the first systems
(A)
(B)
i4 15 G
(C)
FIGURE 9-25 EPR spectra of the di-t-butyl nitroxide radical in

C2H5OH at 25' C. (A) 10-4 M; (B) 10-2 M; (C) 10-1 M.
investigated involved the electron exchange between naphthalene and the

naphthalene negative ion. A second-order electron transfer rate constant of
6 x 10? 1 mole-' sec- was found(2 say in the solvent THF. This is a factor of
100 slower than the diffusion-controlled rate constant; it is thought to be slower
because the positive counter-ion of the negative ion radical in the ion pair must
also be transferred with the electron.
Many effects cause the line widths of one band to differ from those of another
in a given spectrum. We mentioned earlier that the spin density in the CH 3
proton of ethylamine is conformation dependent. The time dependency of this
type of process can influence the line widths of different protons in a molecule
differently. Rapid interchange between various configurations of an ion pair with
an anion or cation radical can also lead to greater line broadening of certain
2
resonances than of others.( 5b,26)
9-11 THE SPIN HAMILTONIAN
The spin Hamiltonian operates only on the spin variables and describes the
different interactions that exist in systems containing unpaired electrons. It can
be thought of as a shorthand way of representing the interactions described
above. The epr spin Hamiltonian for an ion in an axially symmetric field (e.g.,
tetragonal or trigonal) is:
[= - S(S + 1) + g1 IHZ$ + gjfp(H S3 + H,3,) (9-39)
+ A $Z, + Aj($ Z+ $,Z,) + Q, iZ - 1(1 + 1)] gN#NHO
The first term describes the zero-field splitting, the next two terms describe
the effect of the magnetic field on the spin degeneracy remaining after zero-field
splitting, the terms A 1 and A, measure the hyperfine splitting parallel and
perpendicular to the unique axis, and Q' measures the changes in the spectrum
produced by the quadrupole interaction. All of these effects have been discussed
previously. The final term takes into account the fact that the nuclear magnetic
moment pN can interact directly with the external field pNo = gNflNHOI. This
interaction can affect the paramagnetic resonance only when the unpaired
electrons are coupled to the nucleus by nuclear hyperfine or quadrupole
interactions. Even when such coupling occurs, the effect is often negligible in
comparison with the other terms.
In the case of a distortion of lower symmetry, the principal g-values become
g, gy, and g.; the hyperfine coupling constants become A., A,, and A.; and two
additional terms need to be included, i.e., E($3 2 -- 5 ) as an additional zero-field
splitting and Q"(. 2 - 1, 2 ) as a further quadrupole interaction. The symbols P
and P' are often employed for Q' and Q", respectively.
The importance of the spin Hamiltonian is that it provides a standard
phenomenological way in which the epr spectrum can be described in terms of
a small number of constants. Once the values for the constants have been
determined from experiment, calculations relating these parameters back to the
electronic configurations and the energy states of the ion are possible. It should
be pointed out that not all terms in equation (9-39) are of importance for any
given system. For a nucleus with no spin, all terms containing I are zero. In the
absence of zero-field splitting, the first term is equal to zero.
9-12 MISCELLANEOUS APPLICATIONS
When the epr spectrum for CuSiF,- 6H 2 0, diluted with the corresponding
diamagnetic Zn salt, was obtained at 90 K, the spectrum was found to consist
of one band with partially resolved hyperfine structure and a nearly isotropic
g-value( 2 1) In a cubic field, the ground state of Cu2 + is orbitally doubly degenerate.
Although CuSiF 6 6H 2O has trigonal rather than cubic symmetry, this orbital
degeneracy is not destroyed. Thus, Jahn-Teller distortion will occur. However,
there are three distortions with the same energy that will resolve the orbital
degeneracy. These are three tetragonal distortions with mutually perpendicular
axes (elongation along the three axes connecting trans ligands). As a result, three
epr transitions are expected, one for each species. Since only one transition was
found, it was proposed that the crystal field resonates among the three
distortions.(28) When the temperature is lowered, the spectrum becomes aniso-
tropic and consists of three sets of lines corresponding to three different copper
29
ions distorted by three different tetragonal distortions. ' The transition takes
place between 50 and 12 K; the three perpendicular tetragonal axes form the
edges of a unit rectangular solid and the trigonal axis is the body diagonal.
Other mixed copper salts have been found to undergo similar transitions:
(Cu, Mg)3La 2 (NO 3)1 2 -24D 20 between 33 and 45 K, and (Zn, Cu)(BrO 3)2 -6H 2 0,
incomplete below 7 K. The following parameters were reported for
CuSiF, 6H 2 130 ):
90K 20K
g = 2.221 + 0.005 2.46 + 0.01

g = 2.230 + 0.005 g-,= 2.10 + 0.01
A = 0.0021 + 0.0005 cm - g 2.10 + 0.01
B = 0.0028 + 0.0005 cm' A = 0.0110 + 0.0003 cm 1
A} <0.0030 cm
A,
No quadrupole interaction was resolved. A similar behavior (i.e., a resonating

crystal field at elevated temperatures) was detected in the spectra of3 some tris
1
complexes of copper(II) with 2,2'-dipyridine and 1,10-phenanthroline )
The epr spectrum of the complex [(NH 3)Co-O-O-Co(NH 3)1"+is an

interesting example to demonstrate how structural information can be derived
from spin density information and from hyperfine splitting. This complex can be
formulated as: (1) two cobalt(III) atoms connected by an 02 bridge; (2) cobalt(III)
and cobalt(IV) atoms connected by a peroxy, 022 -, bridge; (3) two equivalent
cobalt atoms, owing to equal interaction of one unpaired electron with both
cobalt atoms; (4) interaction of the electron with both cobalt atoms, but more
with one than, the other.
If (1) were the structure, a single line would result, whereas (2) would give
rise to eight lines (I = 7/2 for Co). Structure (3) would result in 15 lines and (4)
in 64. It was found13 2 ) that the spectrum consists of 15 lines, eliminating the
unlikely structures (2) and (4) and supporting structures (1) or (3) or a mixture
of both. An "O hyperfine result would be required to determine the importance
of structure (1).
A study 3 3 ) of the 1: 1 adducts of cobalt(II) complexes with dioxygen has led
to an internally consistent interpretation of the "0 and 59Co isotropic and
anisotropic hyperfine coupling constants. Depending on the ligands attached to
cobalt, the adducts are described as consisting of bound 02 or 02 -
One of the advantages of epr is its extreme sensitivity to very small amounts
of paramagnetic materials. For example, under favorable conditions a signal for
diphenylpicrylhydrazyl (DPPH) radical can be detected if there is 10- gram
of material in the spectrometer. This great sensitivity has been exploited in a
study of the radicals formed by heating sulfur. When sulfur is heated, the
diamagnetic S, ring is cleaved to produce high molecular weight S, chains that
have one unpaired electron at each end. The chains are so long that the
concentration of radicals is low, and paramagnetism cannot be detected with a
Gouy balance. An epr signal was detected,"') and the number of unpaired
electrons (which is proportional to the area under the absorption curve) was
determined by comparing the area of this peak with the area of a peak resulting
from a known concentration of added radicals from DPPH. The total number
of radicals in the system is thus determined, and since the total amount of sulfur
used is also known, the average molecular weight of the species -SSS- can be
calculated. The radical concentration at 300 C was 1.1 x 10 - 'M, and the average
chain length at 171'C was 1.5 x 106 atoms. By studying the radical concentration
as a function of temperature, a heat of dissociation of the S-S bond of 33.4 kcal
mole-' per bond was obtained.
Copper(II) forms complexes of varying geometries, which have similar
electronic spectra and magnetic susceptibilities. Thus, it is often difficult to infer
the geometries of these materials in solution or in media other than the solid,
where single crystal x-ray studies can be used. A recent "0 study of the
five-coordinate adducts formed by various Lewis bases and hexa-
fluoroacetylacetonate copper(II) describes(3 5) a procedure for determining
whether apical or basal isomers of a square pyramid are formed.
This high sensitivity of epr measurements has been of great practical utility
in biological systems. 361 ) Many metalloproteins have been studied in order to
determine the metal's oxidation state, the coordination number of the metal, and
the kinds of ligands attached. The measurements are generally made on frozen
solutions. The interpretation of the results is difficult, and conclusions are based
upon analogies between these spectra and those of model compounds. These
applications are more appropriately considered in the chapter on the epr spectra
of transition metal ions.
There have been several studies reported in which epr has been employed
to identify and provide structural information about radicals generated with high
energy radiation.13 The materials 02 , CIO, C10 2 , P0 32 -, and ClO 3 are among
the many interesting radicals produced by this technique.
Spin traps have been used(3s) to detect radicals whose lifetimes or relaxation
times are too short to be seen by epr. In a typical experiment, a nitrone or nitroso
compound is added to the system in which the radicals are formed. The reactions
in equations (9-40) and (9-41) occur to generate a relatively stable nitroxide that
can be detected in the epr.
H' +R - R (9-40)
N N
101 10.
IO -
(CH 3 ) 3C-NO + R - * (CH 3) 3C-N (9-41)

CH 3
The epr of the resulting nitroxides show the characteristic 1:1:1 nitrogen hyperfine
pattern. In the nitrone, (5,5-dimethyl- I-pyrroline- 1-oxide, equation (9-40)) further
splitting by the hydrogen labeled H' in equation (9-40) gives rise to a six-line
pattern. Splitting by the methyl group of (CH 3) 3CN(O)CH 3 gives an overlapping
triplet of quartets. The magnitude of the AN and AH values can be used to help
identify the radicals trapped. Table 9-3 lists some typical values for spin adducts
of 5,5-dimethyl- I -pyrroline-1-oxide.
TABLE 9-3. Hyperfine Splitting (in gauss) for Nitroxides Produced from
Radical Trapping by 5,5-Dimethyl-1-pyrroline-1-oxide
R. aN aH a,
1. Methyl 14.31 20.52

2. Ethyl 14.20 20.49
3. n-Butyl 14.24 20.41
4. Benzyl 14.16 20.66
5. 1-Phenylethyl 14.20 20.49
6. Phenyl 13.76 19.22
7. x-Cyanobenzyl 14.39 20.63
8. Phenoxymethyl 13.79 19.56
9. 1-Ethoxyethyl 14.20 20.49
10. Tetrahydrofuranyl 14.12 17.92
11. Hydroxymethyl 14.66 20.67
12. 1-Hydroxyethyl 15.03 22.53
13. 1-Hydroxybutyl 14.89 22.72
14. 2-Hydroxypropyl 14.58 23.9,f
15. Acetyl 14.03 17.87
16. Benzoyl 13.99 15.57
17. Aminoformyl 15.23 18.56
18. Dimethylaminoformyl 14.30 17.37
19. Trifluoromethyl 13.22 15.54 1.01(3F)
20. Methoxy 13.58 7.61 1.85(1H)
21. Ethoxy 13.22 6.96 1.89(1H)
22. n-Butoxy 13.61 6.83 2.06(1H)
23. t-Butoxy 13.11 7.93 1.97(1H)
24. Benzoyloxy 12.24 9.63 0.87(2H)
25. Superoxide 12.9 6.9
The couplings are solvent dependent. Oxidation of the nitrone produces a radical with g = 2.009
and AN = 13 gauss that can be confused with trapping a radical. Generation and trapping of radicals
from solvent oxidation must also be considered. A data base, computer compilation of radicals formed
from various traps in a variety of solvents is available.""
REFERENCES 1. F. J. Adrian et al., Adv. Chem. Ser., 36, 50 (1960).

2. A. Carrington, Quart. Rev., 17, 67 (1963).
CITED 3. A. H. Maki and B. R. McGarvey. J. Chem. Phys., 29, 35 (1958).
4. C. E. Strouse and L. F. Dahl, J. Amer. Chem. Soc., 93, 6032 (1971).
5. M. J. S. Dewar, "Molecular Orbital Theory of Organic Chemistry," McGraw-Hill, New
York, 1969.
6. R. W. Fessenden and R. H. Schuler, J. Chem. Phys., 39, 2147 (1963).
7. R. S. Drago and H. Petersen, Jr., J. Amer. Chem. Soc., 89, 3978 (1967).
8. A. D. McLachlan, Mol. Phys., 3, 233 (1960).
9. J. A. Pople et al., J. Amer. Chem. Soc., 90, 4201 (1968) and references therein; J. A.
Pople and D. L. Beveridge, "Approximate Molecular Orbital Theory," McGraw-
Hill, New York, 1970.
10. M. Karplus and G. K. Fraenkel, J. Chem. Phys., 35, 1312 (1961).
11. a. R. E. Cramer and R. S. Drago, J. Amer. Chem. Soc., 90, 4790 (1968).
b. P. D. Sullivan, J. Amer. Chem. Soc., 90, 3618 (1968).
12. M. Broze, Z. Luz and B. L. Silver, J. Chem. Phys., 46, 4891 (1967).
13. I. Biddles and A. Hudson, Mol. Phys., 25, 707 (1973).
14. a. R. S. Drago and H. Petersen, Jr., J. Amer. Chem. Soc., 89, 5774 (1967).
b. R. E. Cramer and R. S. Drago, J. Chem. Phys., 51, 464 (1969).
15. R. W. Fessenden and P. Neta, Chem. Phys. Letters, 18, 14 (1973).
16. R. Fitzgerald and R. S. Drago, J. Amer. Chem. Soc., 90, 2523 (1968).
17. a. R. Livingston and H. Zeldes, J. Chem. Phys., 41, 4011 (1964).
b. J. C. W. Chien and L. C. Dickerson, Proc. Nat]. Acad. Sci. USA, 69, 2783 (1972).
18. H. M. McConnell and R. W. Fessenden, J. Chem Phys., 31, 1688 (1959).
19. T. Cole and C. Heller, J. Chem. Phys., 34, 1085 (1961).
20. H. M. McConnell et al., J. Amer. Chem. Soc., 82, 766 (1960).
21. R. W. Fessenden and R. H. Schuler, J. Chem. Phys., 43, 2704 (1965).
22. P. W. Atkins and M. C. R. Symons, J. Chem. Soc., 4794 (1962).
23. N. Hirota, C. A. Hutchison, Jr. and P. Palmer, J. Chem. Phys., 40, 3717 (1964) and
references therein.
24. H. So and R. L. Belford, J. Amer. Chem Soc., 91, 2392 (1969); R. L. Belford, D. T.
Huang and H. So, Chem. Phys. Letters, 14, 592 (1972) and references therein.
25. a. R. L. Ward and S. I. Weissman, J. Amer. Chem. Soc., 79, 2086 (1957); T. A. Miller
and R. N. Adams, J. Amer. Chem. Soc., 88, 5713 (1966).
b. J. R. Bolton and A. Carrington, Mol. Phys., 5, 161 (1962).
26. J. H. Freed and G. K. Fraenkel, J. Chem. Phys., 37, 1156 (1962).
27. B. Bleaney and D. J. E. Ingram, Proc. Phys. Soc., 63, 408 (1950).
28. A. Abragam and M. H. L. Pryce, Proc. Phys. Soc., 63, 409 (1950).
29. B. Bleaney and K. D. Bowers, Proc. Phys. Soc., 65, 667 (1952).
30. B. Bleaney, K. D. Bowers and R. S. Trenam, Proc. Roy. Soc. (London) A228, 157 (1955).
31. H. C. Allen. Jr., G. F. Kokoszka and R. G. Inskeep, J.Amer. Chem. Soc., 86, 1023 (1964).
32. E. A. V. Ebeworth and J. A. Weil, J. Phys. Chem., 63, 1890 (1959).
33. B. Tovrog, D. Kitko and R. S. Drago, J. Amer. Chem Soc., 98, 5144 (1976).
34. D. M. Gardner and G. K. Fraenkel, J. Amer. Chem. Soc., 78, 3279 (1956).
35. D. McMillin, R. S. Drago and J. A. Nusz, J. Amer. Chem. Soc., 98, 3120 (1976).
36. See, for example, C. S. Yang and F. M. Heuennekens, Biochemistry, 9, 2127 (1970).
37. P. W. Atkins and M. C. R. Symons, "The Structure of Inorganic Radicals," Elsevier
Publishing Co., Amsterdam, 1967.
38. a. E. G. Janzen, Free Radicals Biol., 4, 115 (1980).
b. M. J. Perkins, Adv. Phys. Org. Chem., 17, 1 (1980).
c. E. Finkelstein, G. M. Rosen and E. J. Rauckman, J. Am. Chem. Soc., 102,4994(1980).
d. G. M. Rosen and E. J. Rauckman, Proc. Nat]. Acad. Sci. U.S.A., 78, 7346 (1981).
e. E. G. Janzen, C. A. Evans and J. I.-P. Liu, J. Mag. Res., 9, 510 (1973).
f. D. E. Hamilton, R. S. Drago and J. Telser, J. Am. Chem. Soc., 106, 5353 (1984).
Exercises 401
39. A. S. W. Li, A. H. de Haas, L. J. Park, M. S. Watson and C. F. Chignell, "Spin Trap

Data Base II (STDBII)" MD1703, LMB, P.O. Box 12233, Research Triangle Park,
N.C. 27709.
J. E. Wertz and J. R. Bolton, "Electron Spin Resonance," McGraw-Hill, New York, 1972. ADDITIONAL
C. P. Poole and H. A. Farach, "The Theory of Magnetic Resonance," Wiley-Interscience,
New York, 1972.
REFERENCES*
A. Carrington and A. D. McLachlan, "Introduction to Magnetic Resonance," Harper &
Row, New York, 1967.
C. P. Poole, Jr., "Electron Spin Resonance -A Comprehensive Treatise on Experimental
Techniques," Interscience Publishers, New York, 1967.
B. A. Goodman and J. B. Raynor, Adv. Inorg. Chem. Radiochem., 13, 135 (1970).
W. Weltner, "Magnetic Atoms in Molecules," Dover Publications, Mineola, N.Y., 1989.
A. Carrington, Quart Rev., 17, 67 (1963).
N. M. Atherton, "Electron Spin Resonance," Halsted Press, London, 1973.
P. W. Atkins and M. C. R. Symons, "The Structure of Inorganic Radicals," Elsevier
C. P. Slichter, "Principles of Magnetic Resonance," Harper & Row, New York, 1963.
R. M. Golding, "Applied Wave Mechanics," Van Nostrand Reinhold, New York 1964.
A. Abragam, "The Principles of Nuclear Magnetism," Clarendon Press, Oxford, 1961.
G. E. Pake and T. L. Estle, "Paramagnetic Resonance," 2nd ed., W. A. Benjamin, New
York, 1973.
H. Fisher, "Magnetic Properties of Free Radicals," Landolt-Bernstein Tables, New Series
Group II, Vol. I, Springer-Verlag, Berlin, 1965.
1. Convert the derivative curves below to absorption curves: EXERCISES
(A) (B)
2. a. How many hyperfine peaks would be expected from delocalization of the odd
electron in dibenzene chromium cation onto the rings?
b. Using a procedure similar to that in Fig. 9-6, explain how the number of peaks
arises and what the relative intensities would be.
* For references heavily oriented toward the epr of transition metal ions, see Chapter 13.
3. a. Copper(II) acetate is a dimer, and the two copper atoms are strongly interacting.
The epr spectrum consists of seven lines with intensity ratios 1:2:3:4: 3: 2:1.
Copper nuclei have an I value of 3/2, and copper acetate consists of a ground state
that is a singlet and an excited state that is a triplet. Explain the number and relative
intensity of the lines in the spectrum. [For answer, see B. Bleaney and K. D. Bowers,
Proc. Roy. Soc. (London), A214, 451 (1952).]
b. What would you expect to happen to the signal intensity as a sample of copper
acetate is cooled? Why?
4. Predict the epr spectrum for (S0 3 )2 NO2-
H H
5. The mono negative ion [O O- can be prepared.
-H H
a. How many lines are expected in the spectrum, and what would be the relative
intensities of these?
b. What evidence would you employ and what experiments could be carried out to
indicate electron delocalization onto the oxygen?
c. The magnitude of aH in this material is 2.37 gauss. Compare the spin density on
hydrogen in this molecule with that on a hydrogen atom.
d. How would the sign of the proton hyperfine coupling constant indicate whether
the odd electron was in a sigma- or pi-molecular orbital?
e. Using the value of a, given above and the fact that the unpaired electron is in the
n-system, calculate the spin density on the nearest neighbor carbon.
6. The 1 3C hyperfine coupling in the methyl radical is 41 gauss, and the proton hyperfine
coupling is 23 gauss. Sketch the spectrum expected for 1 3 -CH 3 radical. [For answer,
see T. Cole et al., Mol. Phys., 1, 406 (1958).]
7. Assume that all hyperfine lines can be resolved and sketch the spectrum for the
chlorobenzene anion radical.
8. Assuming all other factors constant, would line broadening be greater for a bimolecular
process with a rate constant of 107 or with a rate constant of 1010?
9. How many lines would you expect in the epr spectrum of (CN) CoO Co(NH ) ?
5 2 35
Explain.
10. The spectrum below is obtained for the NH 2 radical:
a. Convert it to an absorption spectrum.
b. How could you determine whether the larger or smaller splitting is due to hydrogen?
c. Assume that the larger splitting is due to nitrogen. Construct a diagram similar to
Fig. 9-10 to explain the spectrum.
Exercises 403
11. a. How many lines would you expect in the spectrum of the hypothetical molecule
SCI, (I for S = 0 and Cl = 3/2)?
b. Using a procedure similar to that in Fig. 9-6 and Fig. 9-7, explain how this number
arises and indicate the transitions with arrows. State what the expected relative
intensities would be.
12. The epr spectrum of the cyclopentadiene radical (CH,-) rapidly rotating in a single
crystal of cyclopentadiene is given below.
- 30 G
a. Write the appropriate spin Hamiltonian.
b. Interpret the spectrum.
13. Interpret the epr spectrum of -CH 2 OH given below.

1 17.4 G-l
1.15 G--H J--
14. The epr spectrum of CHGe(CH3 )3 ~ is given below. Interpret this spectrum, given
the fact that all of the splittings arise from the phenyl ring protons. Calculate the
a-values.
[- 3 G-i
2
15. Given below is the epr spectrum of
Write the spin Hamiltonian, interpret the spectrum, and report the a-values.
3 G--
Exercises 405
16. a. Interpret the epr spectrum given below and calculate the a-value(s) for the
substituted nitrosyl nitroxide, O
(CH 3)2 N+
C
C-C H5
C
(CH 3 )2 N.
O 10 G---
b. What can you conclude about the delocalization of the unpaired electron?
3' 2' 2 3
17. The epr spectrum of the potassium salt of the biphenyl anion, 4' 4]
is given as follows. Interpret the spectrum and calculate the a-values.
l- 5 G-
18. The epr spectrum of the pyrazine anion, N is given as follows. Interpret
the spectrum and calculate the a-value(s).
1- G-|
19. Below is the epr spectrum of a sample of S2 that has 40% 3 2 S (I = 0) nuclei and
60% 33S (I = %) nuclei. Interpret the spectrum and determine a for 33S.
1
1F000
18501 2050 gauss
20. McConnell's relation allows a rough prediction of the magnitude of proton hyperfine
coupling constants on conjugated organic systens by performing a Huckel m.o.
calculation on the system. The hyperfine constant for the ith proton, ai, is given by
a, = Qpi, where p, = C1 2 .C1j is the coefficient of the various carbon 2p.-atomic orbitals
in the molecular orbital containing the unpaired electron.
a. The carbon 2p, atomic orbitals that make up the it-system are orthogonal to the
C-H sp 2 -sigma bond. Why then, does any unpaired electron density reside on
the proton?
b. The molecular orbital scheme for benzene is:

Exercises 407
16
-i- -i 4-1/
2 05
In the benzene anion, the unpaired electron can be in either 04 or 0s. A Huckel
m.o. calculation for these wave functions gives 60>2
5 3
4
s =- (2<p, - <P2 - (P6 + 2<P4 - <P - (6)

,12
In p-xylene, the degeneracy of these two m.o.'s is lifted, with 14 lower in energy. Use
McConnell's relation and calculate the proton hyperfine coupling constants for the
p-xylene anion. Draw the epr spectrum.
21. In an anisotropic single crystal epr study at v = 9.520 GHz, the g-value was found
to change with rotation in the xz(----- curve)-, yz( .... curve)-, and xy(- curve)-
planes shown as follows.
0 (degrees)
The field position for resonance, H, is given. This is converted to a g-value using
hv 6801.9
AE = hv and equation (9-2). One obtains gd - = H for a 9.520-GHz
microwave frequency. #H H
2
a. Interpolate from the plot and evaluate all of the elements of the g -tensor.
2
b. What would you learn by diagonalizing the g -tensor?
2
c. What steps are required to diagonalize the g -tensor?
d. What steps are required to obtain the direction cosine matrix?
e. Write the spin Hamiltonian.

The Electronic Structure
and Spectra of
Transition Metal Ions
10
The subject of this chapter has also been the topic of several textbooks."' 12) Introduction
Here we shall present an overview of the electronic structure of transition metal
ions. In so doing, we will develop some important ideas for the understanding
of the spectroscopy of transition metal ion complexes-our main objective.
Transition metal ion systems are covered separately in this book because their
unpaired electrons introduce several complications. As is so often the case, these
complicating factors, when understood, provide a wealth of information. The
complications involve electron-electron interactions, spin-orbit coupling, and the
influence that a magnetic field has on systems with unpaired electrons. We have
discussed many of these topics earlier, but their full implication is best demon-
strated with examples from transition metal ion chemistry.
10-1 ELECTRON-ELECTRON INTERACTIONS AND Free Ion

TERM SYMBOLS Electronic
There are numerous ways in which one or more electrons can be arranged in States
the five d orbitals of a gaseous metal ion. We can indicate the energy differences
arising from different interelectronic repulsions and different orbital angular
momenta for these various arrangements with term symbols. Any one term symbol
groups together all of the degenerate arrangements in the gaseous ion. The
simplest case to consider first is d'. There are five ways to arrange an electron
with m, = + /2 in the five d orbitals. Each arrangement is called a microstate
configuration.
2 1 0 -1 -2
In the absence of external electric and magnetic fields, the five microstates are
degenerate, and there are five others that are also degenerate with these,
corresponding to m, = - /2. These ten microstatescomprise the tenfold degeneracy
of the so-called 2 D term (vide infra).
410 Chapter 10 The Electronic Structure and Spectra of Transition Metal Ions
The ground state term for any d" configuration can be deduced by arranging
the electrons in the d orbitals, filling those with the largest m, values first and
not pairing up any electrons until each orbital has at least one; i.e., Hund's rules
are obeyed. The m, values of the orbitals containing electrons may be algebraically
summed to produce the L value for the term. More completely, the m/ quantum
number for an individual electron is related to a vector with component m,(h/27)
in the direction of an applied field. The ML value is the sum of the one-electron
m/ values. Vector coupling rules demand that M, have values of L, (L - 1), ... ,
-L, so we deduce that the maximum ML value is given by the value of L. The
following letters are used to indicate the L values: S, P, D, F, G, H, I corresponding
to L = 0, 1, 2, 3, 4, 5, 6, respectively. The 2L + 1 arrangements refer to the orbital
degeneracy and are described by ML with values of L, (L - 1),..., -L. As
mentioned earlier, the 2D term symbol describes the d' case. It is tenfold
degenerate with a fivefold orbital degeneracy corresponding to ML values of 2,
1, 0, -1, -2. In the d' ion, the ground 2 D term is the only one arising from the
3d orbitals.
The spin multiplicity of a state is defined by 2S + 1 (S, in analogy with L,
is the largest possible Ms, where Ms = Xms) and is indicated by the superscript
to the upper left of the term symbol. The spin multiplicity refers to the number
of possible projections of S along a magnetic field; e.g., when S = 1, the
multiplicity of three refers to Ms = 1, 0, - 1 (giving the z-component of spin
angular momentum aligned with, perpendicular to, and opposed to the field).
The total degeneracy of a term is given by (2L + 1)(2S + 1). The value of the S
quantum number for the term (or state) is given by the maximum M5, which
equals the sum of the m, values of all unpaired electrons. Complete subshells
contribute nothing to L or S, because the sum of the m, and the m1 values is zero.
Next, consider the d 2 configuration. There are 45 ways to arrange two
electrons with m, = t 1/2 in the five d orbitals. Using the procedures described
above for the microstate I1 1 7 j, we see that L = 3 and S = 1.This

leads to a 3F ground term that is 21-fold degenerate in the absence of spin-orbit
coupling. The other 24 microstates comprise higher energy (excited) states; i.e.,
electron-electron repulsions are larger for these states. All of the terms for a d"
configuration can be found by constructing a table like that shown in Table 10-1
for a d2 ion. To be systematic, we can begin construction of the table with the
row ML = 4. This ML value can be obtained only by having two electrons in
m/ = 2, and this can be done only if the spins are paired up. The resulting
microstate is abbreviated as 2 2 -. The microstate is placed in the M, = 0 column.
Next, ways to obtain M, = 3 are shown. The possibilities are 2+ 1+ , 2+1 -, 2 1',
and 2 -1-, corresponding to Ms values of + 1, 0, 0, and - 1, respectively. The
procedure is repeated for ML values of 2, 1, and 0. The microstates corresponding
to negative M, values are not indicated in the table. They are obtained by
multiplying the ML values above for the positive M, microstates by -1; e.g., for
ML= -3, the possibilities are -2+ 1+, -2+ - 1-, -2- - 1+, and
-2- - 1-. Note that, for example, 2 + 1 is not distinct from 1- 2 + since the
order in which we list the electrons is irrelevant; but it is distinct from 2- 1 +
because in this latter microstate the electron with m, = 2 no longer has m, = + '/2-
Starting with the highest ML value, we can conclude that there must be a
G term or state and that it has ML components of 4, 3, 2, 1, 0, -1, -2, 3,
-4. A box is used to set these configurations apart. The choice of the ML = 3
10-1 Electron-Electron Interactions and Term Symbols 411
TABLE 10-1. Microstates for a d2 Ion for Positive ML Values

Ms
M, +1(11) 0(±) 1()
2*2-
(2+1 ) 21 (2-1*) (2-1-)
(2*0*) 2*0 (2- 0) (2-0-)
(2+ - 1) 2 1 (2- - 1) (2- - 1-)[1 -0-1
[1*0+] [1 0*]
(2* - 2*)[1* - ] 2 2 (2 (2~ - 2-)[1 - 1-]
1 I[1
0~-0
configuration for the 'G term is arbitrary because we are only bookkeeping with
this procedure. The actual wave function for this component of the 'G term is a
linear combination* of the two microstates indicated for ML = 3.,Raising and
lowering operators can be employedm) to produce the wave function for given
values of L, ML, S, and Ms. The same is true whenever there is a choice of
microstates. Now, we proceed to the next highest ML value that remains, namely,
ML = 3. With Ms = +1, 0, -1 components, it can be deduced that there must
be a 3F term. This term will be 21-fold degenerate; the 12 microstates, with
non-negative ML values, arbitrarily assigned to it are enclosed with parentheses
in Table 10-1. Next, we come to a state with L = 2, which must be a singlet (i.e.,
S = 0). The microstates arbitrarily assigned to this 'D term are circled. Next, we
enclose with brackets those microstates of the 3P term. The remaining term is
'S. Each of these terms constitutes a degenerate set of states, and each term differs
in energy from any other.t
* Each of the microstates is really an abbreviation for a determinantal wave function, e.g.,
1 2 (1) 1-(1)
2 (2) 1 (2)l
t One often finds the individual eigenfunctions of a term referred to as states. The entire term
is also referred to as a state. The difference is usually obvious. If it is not, we shall use term or level
to describe the entire collection of degenerate states and component states for the individual states.
The energies for all of these terms can be calculatedm) and expressed with
the Condon-Shortley parameters FO, F 2 , and F4 . These parameters are abbrevi-
ations for the various electron repulsion integrals of the ion. The energy expression
for any term as a function of these parameters is independent of the metal ion.
The magnitude of the parameters, on the other hand, varies with the metal ion.
For example, E( 3 F) = FO - 8F 2 - 9F4 and E(3 P) = F, + 7F 2 - 84F 4 . The tran-
sition energy 3P _ 3 F is the difference between the energies of these two terms
of 15F 2 - 75F 4 . Similar expressions exist for all the other transitions involving
terms of the gaseous ion. The entire spectrum can be fit with the parameters F
2
and F 4 . This is true for any d 2 ion.
In the V(III) ion, the 3F-3P transition occurs at 13,000 cm and the 3 F-'D
transition occurs at 10,600 cm -. Solving the two simultaneous equations
E(3 P _ 'F) = 15F 2 - 75F 4 = 13,000 cm 1 (10-1)

E('D _ 3F) = 5F 2 + 45F 4 = 10,600 cm -1 (10-2)
'S
one obtains F 2 = 1310 cm- I and F 4 = 90 cm- '.
1G The Racah parameter redefined the empirical Condon-Shortley parameters
so that the separation between states having the maximum multiplicity is a
function of only a single parameter, B:
(9)P
(5)
B = F 2 - 5F4 (10-3)
,
A second parameter, C, is needed to express the energy difference between terms

(21 ) of different multiplicity:
//,e-- 3
F
C = 35F 4 (10-4)
No Electronic Electronic
Interactions Interactions Rearranging, we obtain F 2 = B + C/7 and F4 = C/35. Substituting these into
(5 degenerate equations (10-1) and (10-2), we obtain:
d orbitals)
FIGURE 10-1 The terms E(3P -- 3F) = 15B
arising from the
electron-electron interactions E('D _ 3F) = 5B + 2C
in a d2 gaseous ion. The
number in parentheses For V(III), we find B = 866 cm 1 and C/B = 3.6.
indicates the degeneracy of In summary, the results of the electron-electron interactions in a d 2 ion give
each level (excluding any rise to a 'F ground term and the excited states arising from the d orbitals shown
spin-orbit coupling). in Fig. 10-1. The degeneracy of each term is indicated in parentheses.
TABLE 10-2. Free Ion Terms for Various d4 Ions

n Terms
d' d9 2D
d2 d8 3
F 3
P 1G 'D IS
d3 d7 F4 P 2
H G 2F 2 D 2 D 2p
2
3
d4 d6 5D H 3G 3 3 F 3 D 3 P 3p I 1G 1G 1F1 D 1D tS iS
4 4
d5 6S 4G F D 4 P 21 2 H 2 G 2 2F 2 2D 22D 2p 2S
By using procedures similar to those employed in Table 10-1, we can
determine the terms arising from various d" ions. The results for n = I to n = 9
are presented in Table 10-2.
The d9 configuration is for many purposes considered to be equivalent to
the d' case, if we think in terms of the degenerate states that would arise from
degeneracies associated with the positive hole that exists in the d 9 case. It may
help to think of d9 as being a d10 case with a positron that can annihilate any
one of the 10 electrons. This concept is referred to as the holeformalism. By the
same token, the following equivalences arise:
d2 dd8
ds ~ d?7
d'4 d6
10-2 SPIN-ORBIT COUPLING IN FREE IONS
As discussed in Chapter 9 (Fig. 9-18), the coupling of the magnetic dipole from
the electron spin moment with the orbital moment, L - S, is spin-orbit coupling.
Variations in the amount of spin-orbit coupling in the different electronic
configurations also lead to splitting of the terms derived so far. Two schemes are
widely used to deal with this effect: the so-called Russell-Saundersor L - S coupling
scheme, and the j - j coupling scheme. When the electron-electron interactions
give rise to large energy splittings of the terms compared to the splittings from
spin-orbit coupling, the former scheme is used. With the L - S scheme, we
essentially treat the effects of spin-orbit coupling as a perturbation on the
individual term energies. On the other hand, the j - j coupling scheme is used
when a large splitting results from spin-orbit coupling and the electron-electron
interactions are sufficiently small to be treated as a perturbation on the spin-orbit
levels. The j - j scheme is applied to the rare earth elements as well as the third
row transition metal ions. Briefly, in the j j scheme the spin angular momentum
of an individual electron couples with its orbital momentum to give a resultant
angular momentum, j, for that electron. The indivdual j's are coupled to
produce the resultant vector J for the system, labeling the overall angular
momentum for the atom. The L S coupling scheme is applicable to most first
row transition metal ions, and we shall discuss this scheme in more detail. We
previously mentioned that the individual orbital angular momenta of the
electrons, i, couple to produce a resultant angular momentum indicated by L.
The spin moments couple to give S. The resultant angular momentum including
spin-orbit coupling is given by J, and the corresponding quantum number J can
take on all consecutive integer values rangingfrom the absolute values of | L - SI
to |L + S 1. For the ground term, the minimum value of J refers to the lowest
energy state of the manifold if the subshell (e.g., the d orbitals) is less than half
f lled; and the maximum value of J refers to the lowest energy state when the subshell
is more than halffilled. If the shell is half filled, there is only one J value because
L = 0.
Our discussion in this section can be made clearer by working out some
examples. The box diagram for the ground state of the carbon atom is:
2
p
Is 2s +1 0 -1
b U '1 1
The value of the L quantum number, obtained by adding the m/ values for all
the electrons in incomplete orbitals, is 1 for carbon: L = + 1 + 0 = 1. This L
value corresponds to a P state. The sum of the spin quantum numbers (ms = + 1/2)
for all unpaired electrons, is 1 for carbon: S = 1/2 + '/2 = 1. The multiplicity is
three, and the term symbol for the ground state is 3P. The values for J (given
byIL - SI,. . . ,I L + S|)areIL -S 1 - = 0,1L + S = 1 + I = 2,soJ = 0,
1, 2 (one being the only integer needed to complete the series). The subshell
involved is less than half filled, so the state with minimum J has the lowest energy.
The term symbol for the ground state of carbon is 'Po, with the zero subscript
referring to the J value.
The box diagram for the ground state of V" is
+2 +1 0 -1 -2
with term symbol 3F 2 (L = 3, S = 1, J = 4, 3, 2). An excited state for this species
is represented by 1k ; this microstate belongs to the term with
term symbol 'G4 (L = 4, S = 0, J 4). For nitrogen with a box diagram

1 1 1 , L = 0, S = 3/2, and J = '/2 so the term symbol 4S3/2 results.
Note that with L = 0 there is only one J value because IL + S|
IL - SI = 3/2-
For practice, one can determine the following term symbols for the ground
states of the elements in parentheses: 3P2 (S), 2 3/2 (Cl), 3F 2 (Ti), 5D, (Cr), 3F
4
(Ni), 3Po (Si), 4 S3/2 (As), and 'I9/2 (Pr).
Two parameters, and 2, are commonly used to describe the magnitude of
the energy of the spin-orbit coupling interaction. The parameter is used to
describe the spin-orbit coupling energies for a single electron. It measures the
strength of the interaction between the spin and orbital angular momenta of a
single electron of a particular microstate, and is thus a property of the microstate
and not of the term. The operator is (l - S. The value of is given by
Zeffe
= 2m 2 c 2 r -3> (10-5)
where (r-3> is the average value of r-3, m is the mass of the electron, c is the
speed of light, and Zeff is the effective nuclear charge.
When the parameter 2 is used to describe the term, the operator now becomes
AL - S. The values of 2 and are related by
A = + /2S (10-6)
The parameter is fundamentally a positive quantity. If the shell is less than half
filled, the sign of 2 is positive; if it is more than half filled, 2 is negative. This
makes sense if we think in terms of positive holes requiring the sign of equation
(10-5) to change for the configurations of more than half-filled shells. In summary,
then, for a shell that is less than half filled, the lowest value of J corresponds to
the lowest energy and 2 is positive.
An equivalent operator form for L -S is given by 1/2( J2 _ L2 _2) when the
states can be characterized by quantum numbers L, S, and J. The spin-orbit
contribution to the energy of any level is then given by:
A J( J + 1) - L(L + 1) - S(S + 1)] (10-7a)
The energy difference between two adjacent spin-orbit states in a term is given by
AE, + 1 = 2( J + 1) (10-7b)
For example, the energy separation between the J = 3 and J = 4 states of a term
is 42. Furthermore,in the L - S scheme, spin-orbitsplitting occurs so as to preserve
the center of gravity of the energy of the term, i.e., the average energy remains
the same. The ground state for a d2 system, 3F, has J values of 4, 3, and 2, with
2 lowest since the shell is less than half filled. The complete ground state term
symbol is 3 F2 . The 'D excited state has only one possible J state, equal to 2.
is) (1) J =0
IG() - (9) J=
(5) A J 2
3p(9
(3) - A =1
(1) - 2A j 0
FIGURE 10-2 Spin-orbit states from a
3
ID (5) (5) J = 2 d 2 configuration. The splitting of the F2
state by a magnetic field Ho is indicated
(9) 3A J 4 on the far lower right.
(21)'
J3
-- - 2
(5) - 4X J =1
0
~ - 1
-2
e- -e- AL-S H
For the excited state, 1P, we have J = 0, 1, and 2, whereas 'G has only J = 4
and 'S has only J = 0.
Now, using equation (10-7a), we can calculate the spin-orbit contribution
to the energies of all the J states. For the ground level of 3 F where J = 2, we
obtain (1/2)2[2(2 + 1) - 3(3 + 1) - 1(1 + 1)] = -42. This result is summarized
in Fig. 10-2 along with the results of similar calculations of the effects of 2L - S
on all the states of a d 2 system. Not all the degeneracy is removed by spin-orbit
coupling, and the remaining degeneracy, corresponding to integer values of Mj
from J to -J, is indicated in parentheses over each level. Note that equation
(10-7b) is obeyed and the center of gravity is preserved. For example, in the 3P
term, the degeneracy times the energy change gives 52 - 32 - (1) (2)2, = 0. The
degeneracy of the individual J states is removed by a magnetic field. The splitting
into the Mj states is indicated only for the ground J = 2 term in Fig. 10-2.
Crystal Fields 10-3 EFFECTS OF LIGANDS ON THE d ORBITAL

ENERGIES
We usually do not work with gaseous ions but with transition metal ions in
complexes. There are two crystal field type approaches to determine the effects
that these ligands in a transition metal ion complex have on the energies of the
d orbitals. The metal ion electrons in a complex undergo interelectronic repulsions
and are also repelled by the electron density of the Lewis base (ligand). When
the repulsions between the metal electrons and the electron density of the ligands
is small compared to interelectronic repulsions, the so-called weak field approach
is employed. When the ligands are strong Lewis bases, the ligand electron-metal
electron repulsions are larger than the interelectron repulsions and the strong
field approach is employed.
The basis set used in these problems can be the orbitals represented by
complex wave functions whose angular dependences are given by the spherical
harmonics
2
Y2 0 = (5/8)'/2(3 cos 0 - 1). (2n)- 12
Y2 1"= (15/4)' 2 sin 0 cos 0 -(27)- /e'

2
Y 2
= (15/16)2 sin -(20) e
Alternatively, the real trigonometric wave functions, which are linear combina-
tions of the complex orbitals taken to eliminate i, can be employed. These are
given by:
d2 0> (d,2 is really d(z 2 /3))
d,, = (i1')11 - 1> + |1)]

dx, (1/, )[ I - 1) -- [I>]
d ,= -(i/ 2)-12> - 2>]

d2 - =(1/ 2)+ ± - 2>]
10-3 Effects of Ligands on the d Orbital Energies 417
In the weak field approach, the free-ion term state eigenfunctions (which
take into account the interelectronic repulsions in the d-manifold) are employed
as the basis set. As an example, for the 3F term, the wave functions corresponding
to M, = ±3, +2, 1, 0 are used. They are abbreviated as 13>, 12>, etc. The
Hamiltonian is given as:
H=H+f±V
where No is the free-ion Hamiltonian and Vis taken as a perturbation from the
ligand electron density on Ho. The perturbation, V, has a drastically simplified
form incorporating only the electrostatic repulsion from the ligands, which are
represented simply as point charges. For an octahedral complex, the perturbation
is given by:
V= eZi/rij (10-8)
i=1
where e is the charge on the electron, Z is the effective charge on the ith ligand,
and r 1 is the distance from the d-electron (this is a d' problem) to the ith charge.
This is to be compared to the full Hamiltonian given in Chapter 3. Using the
simplified Hamiltonian [equation (10-8)] leads to crystalfield theory. It is to be
emphasized that this formulation of the problem simply describes the electrostatic
repulsion between the d-electrons and the ligand electron density, and as such
can tell us directly only about relative energies of the d orbitals.
In order to evaluate the integrals <M I V IM,'>, 7 is written in a form that
facilitates integration."," When this is done, many quantities related to the radial
part of the matrix elements appear in the secular determinant with the form
'(Ze 2r2 4a-5). Here, r2 - corresponds to the mean fourth-power radius of the
d-electrons of the central ion, a is the metal-ligand distance, and Ze has the same
units as e. This radial quantity is referred to as 10Dq and has units of energy.
It is informative to write the secular determinant for an octahedral complex
with this Hamiltonian acting upon a d' configuration. Employing the complex
d-orbital basis set, we obtain
12> 1) 0> I -1 1-2>

|2> Dq - E 5Dq
|1> -4Dq - E 0
|0> 6Dq - E
I-1> -4Dq - E
S-2> 5Dq Dq - E
This gives roots
E(I1>)= -4Dq
E(I -I>) = -4Dq
E(I0)) = 6Dq
and the determinantal equation
|2> - 2>
|2> Dq - E 5Dq 0
I -2> 5Dq Dq -E
This determinant is solved to produce two energies: one at -4Dq and one at 6Dq.
As in a Hickel calculation, the energies can be substituted into the secular
equations written from the secular determinant, and the wave functions thus
obtained. The results are
04 = -i2- 11(12)> 1-2>)
= 2- + | -2) )(2)
6Dq- These are the wave functions for the d, and d,2_ orbitals, the latter pointing
2
d,, d,2_ 2 at the ligands and the former in between the ligands. Note that the octahedral
E crystal field mixes the |2> and I -2> wave functions and makes it more convenient
N
to employ the real d, and dX2 _,2 orbitals in the description of the complex. Since
E
the degeneracy of I1> and I -1> is not removed, the wave functions 0, to
R 3
G can be considered as the real or imaginary combination as is convenient. These
Y results are summarized in Fig. 10-3, where it can be seen that two degenerate
-4Dq- sets of orbitals result which are separated by 10LDq. Thus, we expect one d-d
die duz d, electronic transition for a d' system with an energy corresponding to 10Dq. In
0
h symmetry, the three degenerate d., d,2, and d, orbitals transform as t2g and
FIGURE 10-3 Splitting of lead to the 2 ground state, while d22 and d,2 _,2 lead to the 2 E. excited state.
the one-electron d orbitals by
The following information is conveyed by these symbols: (1) the symbol T indicates
an Oh crystal field.
that the state is orbitally triply degenerate and E doubly degenerate; (2) the
superscript 2 indicates a spin multiplicity of two, i.e., one unpaired electron; and
(3) the g indicates a gerade or symmetric state.
Next, we shall treat a weak field, octahedral d2 complex. Our initial concern
will be with the 3F term described by the basis set 13> . .. 10> . .. I 3>. Since
our basis set for 3F consists of all Ms = 1 functions, this has been dropped from
the symbol and only ML is indicated; i.e., 13, 1>, etc., would be a more complete
description. We shall not present the wave functions for these basis functions.
Procedures for obtaining these wave functions using the raising and lowering
operators are presented by Ballhausen.1 The secular determinant is given below:
13)> |2> |1) 10) 1) 1-2> 1-3>

(31 -3Dq - E 15' 2 Dq
(21 7Dq- E 5Dq
(1 - Dq - E 15 /2 Dq
<01 -6Dq- E
2
( - II 15' Dq Dq - E
( - 21 5Dq 7Dq - E
( - 31 15' 2 Dq 3Dq - E
The solutions are
E 1 = -6Dq E,=2Dq
E 2 = -6Dq E6 = 2Dq
E,= -6Dq E,= 12Dq
E4 = 2Dq
The energies* and wave functions for the respective levels are indicated in Fig.
10-4.
12Dq [2 1/2 (12> + 1-2))]

3
A2 , (3)
[241/2 (313) + 151/2 _1i))
3
T2 , (9) [24-1/2 (31- 3) + 151/2 1))]
N 3
E F /
E -....p - 2Dq [2-1/2 (12) - 1-2))]
R (21) 3T 9
G T* (9) - 6Dq [24-1/2 (151/2 -3) -- 311))]
y [10>]
[24-1/2 (151/2 |3) -31 -1))]
3 2
FIGURE 10-4 Splitting of the Fterm of a d ion by an Oh
crystal field.
In this analysis, we have ignored any covalency in the metal-ligand bond.

As a result, if we were to attempt a quantitative calculation of Dq, it would differ
considerably from that found experimentally. Ligand field theory admits to
covalency in the bond and treats Dq (and other parameters to be discussed shortly)
as an empirical parameter that is obtained from the electronic spectrum. The
formulation of the problem in all other respects is identical.
10-4 SYMMETRY ASPECTS OF THE d-ORBITAL SPLITTING

BY LIGANDS
As is usually the case when a vastly simplified Hamiltonian is employed, the

correct predictions from the method are symmetry determined. For example, we
mentioned in Chapter 2 that appropriate combinations of the binary products
of the x, y, and z vectors gave the irreducible representations for the d orbitals
and their degeneracies. We can use principles already covered (Chapter 2) to
illustrate a procedure for deriving all of the states arising from one-electron levels.
This procedure can then be extended in a straightforward way to derive all of the
states arising for various multielectron configurations in various geometries.
We begin by examining the effect of an octahedral field on the total
representation for which the set of d wave functions forms the basis. To determine
this total representation, we must find the elements of matrices that express the
* Dq is defined for a one-electron system. For a polyclectron system, one should employ a
different Dq value for each state. (C. J. Ballhausen and H. B. Gray, "Chemistry of Coordination
Compounds," Volume I, ed. A. E. Martell, Van Nostrand Reinhold, Princeton, N.J., 1960.) This is
seldom done in practice because of the crude nature of this model.
effect upon our basis set of d orbitals of each of the symmetry operations in the
group. The characters of these matrices will comprise the representation we seek.
Since all of the d orbitals are gerade, i.e. symmetric to inversion, no new
information will result as a consequence of the inversion symmetry operation.
Thus, we can work with the simpler pure rotational subgroup 0 instead of 0
If you need to convince yourself of this, note that in any group containing i (e.g.,
D 4 h or C 3 ), the corresponding rotation group (e.g., D4 or C 3) has the same
irreducible representation for the binary products except for the u and g subscripts
in the former group. Recall that the d wave functions consist of radial, spin, and
angular (0 and cp) parts. The radial part is neglected, for it is nondirectional and
hence unchanged by any symmetry operation. We shall assume that the spin part
is independent of the orbital part and ignore it for now. The angle 0 is defined
relative to the principal axis (i.e., the rotation axis), so it is unchanged by any
rotation and can be ignored. Only (p will change; the form of this part of the
wave function is given by e'"'-. (For the d orbitals, t = 2 and m, has values 2, 1,
0, - 1, -2.) To work out the effects of a rotation by c on e"", we note that such
a rotation causes the following changes:
e2ipe 2i( p+a)
e'o ei((+ )
rotate
ebya
e-2ie -2i(P+2)
e -2hp e -2(p+a
The matrix that operates on our d-orbital basis set is
Se 2i2 0 0 0 0
0 e' 0 0 0
0 0 eo 0 0
0 0 0 e-' 0
0 0 0 0 e-2ia
A general form of this matrix for the rotation of any set of orbitals is given by:
-e""a 0 ... 0 0
0 e"-lu ~ .... 0 0
0 0 ... e"l-'194o 0
0 0 ... 0 e
When a = 0, each element is obviously one. We give the trace of this latter matrix
without proof* as:
x(0) = sin (10-9)

sin ()
+
* The quantities summed form a geometric progression.

where a # 0. From the trace determined with this formula, we have the character
of the representation for any rotation operation. Substituting directly into this
formula, we find that the total character for the C3 rotation of the five d orbitals
is given by
sin 2+ - sin -sin -

K
C 2 3 3 3
x(C 3 )= - -- * - -
sin sin sin
s3 x 2 s 3 3
Characters for other rotations can be worked out in a similar way. To obtain
the characters when oc= 0 and a = 27r, one must evaluate the limit of an
indeterminate form, e.g.,
sin ( ± (2+)
sin 7 0
Using l'Hopital's rule, one obtains
Z(0) = 2/ + 1
X(27r) = 2/ + 1 for integer / or
-(2/ + 1) for half-integer /
Thus,-for the identity, i = 0 and Z(E) is given by 2/ + 1. Referring to the character
table for the 0 point group and using the above formula to determine the
characters for the various operations on the five d orbitals, we have:
E 6C4 3C2 (= C} 2) 8C3 6C2

XT 5 -1 1 -1 1
Using the decomposition formula, we obtain the result yT = E + T2 . Since the d

orbitals are gerade, we can write this as
XT = E + T2g
This was the result of our crystal field analysis. By a similar symmetry analysis,
the results summarized in Table 10-3 can be obtained.
We could work out the effect of other symmetries on the one-electron levels
in a similar way. Alternatively, one can use a correlation table that shows how
the representations of the group 0 are changed or decomposed into those of its
sub-groups when the symmetry is altered. Table 10-4 contains such information
for some of the symmetries commonly encountered in transition metal ion
complexes.
With the irreducible representations given in Table 10-3 for various atomic
orbitals in 0 , symmetry and with the correlation table given in Table 10-4, we
can ascertain the irreducible representations of the various orbitals in different
symmetry environments. The results for single electrons in various orbitals apply
TABLE 10-3. Representations for Various Orbitals

in Oh Symmetry
Type of
Orbital / Value Irreducible Representation"
S 0 al.
p 1 ti.
d 2 e,+ t2g
f 3 a2,+tl+t2,
g 4 ag+e,+tlg+t2,
h 5 e.+2t,+t2.
i 2
6 aig+a 2 g+eg+tg+ t2g
" The subscripts g and u are determined by the g or u nature of

the atomic orbitals involved. When / is even, the orbital is gerade;
when / is odd, the orbital is ungerade.
TABLE 10-4. Correlation Table for the Oh Point Group
Of 0 Td D4 h Dza C4, C2. D 3d D, C 2h
Aig Al Al A1g Al A1 A1 Aig Al Ag

A2g A2 A2 Big B1 B1 A2 A2g A2 B,
Eg E E Ag+ Big A 1 + B1 A 1 + B, A 1 + A2 Eg E Ag+ Bg
Tg T1 T1 A 2 ,+Eg A2 +E A 2 +E A2 +B 1 +B 2 A 2 g+Eg A 2 +E Ag+ 2 Bg
Tg T2 T2 B 2 g+E, B2 +E B2 +E A 1 +B 1 +B 2
2 Alg+Eg A 1 +E Ag +Bg
Alu A1 A2 Al. B1 A2 A2 Al. A1 Au
A2, A2 A1 Blu A, B2 Al A2 u A2 B.
E, E E A + Bu A 1 + B1 A 2 + B2 A 1 + A2 E E Au+ B.
Tu T T2 A 2 ,+E. B2 +E A 1 +E A 1 +B 1 +B 2 A 2 .+E. A 2 +E Au+2B,
Tu T2 T B 2 u +Eu A2 +E B 1 +E A2 +B 1 +B 2 A 1u+Eu A 1 +E 2A +B,
also to the terms arising from multielectron systems. For example, we can take
the 3 F, 3 p, 1G, 'D, and 'S terms of the d2 configuration and treat them likef,
p, g, d, and s orbitals. The g or u subscripts given in Table 10-3 will not apply,
but will depend upon the g or u nature of the atomic orbitals involved. Thus,
Table 10-3 applies to terms as well as orbitals. The D term, for example, is fivefold
degenerate like the five d orbitals, the former being described by a wave function
for each of the five ML values. These wave functions have a <D part given by e',".
Combining Tables 10-3 and 10-4, we can see that the D state of the free ion
splits into Eg + T2, states in an octahedral field and into Ag + Big + Eg + B2 g
states in a tetragonal D4, field. Similarly, the 3 F term gives rise to A g + Ttg + Tg
2
in an octahedral field and B1 + A 2 + 2E + B 2 in a C,, field.
Next, we shall consider the states that arise from a d2 configuration when
the crystal field is large compared to the interelectronic repulsions.* The
interelectronic repulsions are treated as perturbations on the strong field
d-electron configurations. In other words, the various crystal field states are
identified, eigenfunctions are constructed, and e2 /rg. is used as a perturbation.
The terms are readily written for the various d" configurations. For d', the two
terms are 2 T2, and 2 Eg. The terms that arise from a configuration with an
additional electron, d"',are given by the direct product of the d" term symmetries
and that of the added electron. For d 2 , we have t2g , t2 eg , and e 2 arrangements.
Accordingly, for t2,2, we have
t2, X t2, which leads to T1, + T2 , + Eq + A q
For t2 'e , we have
t2, x e, which leads to T1 , + T q

2
and for e., we have
eq x eq which leads to E9 + Ajg + A2q
Next we have to determine the multiplicity of these terms and show the
connection of these strong terms to those of the gaseous ions. The multiplicities
are determined by the method of descending symmetries." By considering the
eq configuration, we can show that when we lower the symmetry from 0 h to
D4 h to remove the degeneracy, the determination of the spin degeneracy becomes
straightforward. The correlation table (Table 10-4) shows
e aiq
big
Oh D 4h
* The total degeneracy that exists for a system of q electrons filling a subshell with z-degenerate
orbitals (each occupied by two electrons with opposite spin) is given by:
(2z)!
q!(2z - q)!
For a tg2 configuration, we have:
(2 x 3)! = 5
2!(6 - 2)!
For t2,e,', we have
6! 4! 416
4]=6
12! (6 - 2)! 1 !(4 - 1)!] =[5]
In addition to the orbital correlation shown above, it is important to

remember that states behave in the same way. In D 4 h, ag 2 must be singlet because
of the Pauli principle. For example, alg2 leads to aig x ag which leads to 'A ,g
while b g 2 leads to big x big which leads to A ig, and ag'big' leads to big x alg
which leads to 'Big or 3Big. The states in octahedral symmetry (Alg, A 2 g, and
Eg) must change to those in D 4h symmetry as shown in the group correlation
table (Table tO-4) and as summarized below:
O, D 4h
Aig ----Alg
A 2g -- Big
Big
We have determined the multiplicity of the D4, states; since lowering the symmetry
cannot change the spin degeneracies, we can work backwards to determine the
spin degeneracy of the Oh states. The 'Alg state in D4, must correspond to a
singlet 1Alg in 0 , The other 1Alg must arise from Eg, requiring that this be 'Eig.
The Big state in D4, arising from 'Eg must also be a singlet. We are left with
the fact that 3 Big in D 4 . is associated with A 2g, requiring this to be 3A 2 g-
D 4h O,
3
Big A2g( 3A 2g)
gAlg Eg('Eg)
'Big
The t2g2 configuration gives rise to Ajg + Eg + Tig + T2g states. We must
examine the correlation table for these states in Oh symmetry and find a lower
symmetry that converts these states to one-dimensional representations or a sum
of one-dimensional representations. Table 10-4 shows that C 2h and C 2 , satisfy
this requirement. The results for C 2h are summarized as
Oh C 2h
A, g ---- Ag
Eg Ag + Bg
Tig A + B + Bg
T2, Ag + Ag + B
Since t2, in O, gives rise to a. + a. + bg orbitals in C2 , the possible configurations

are a. 2, agi ag2 1 , a, b, 1,ag2 2, ag2 b, b 2 . As previously:
ag, x ag1 leads to Ag
ag1 x ag leads to Ag
ag1 x bg leads to B9
ag2 x a8 2 leads to A.
ag2 x b, leads to B9
bg x b. leads to A.
The subscripts 1 and 2 on the a, orbital are used to distinguish the two different TABLE 10-5. Relation
ag representations that result from t 2 , in this lower symmetry. Since the first, of the Multiplicities in
fourth and sixth binary products listed correspond to both electrons occupying C2, and Oh
the same orbital* (a81 , ag2 , and b, respectively), these must be singlet states: 1A 8, 0
'Ag, and 'A., respectively. The second, third, and fifth binary products correspond c2h h
to electrons in different orbitals and give rise to singlet and triplet states: iAg + 3 Ag
and 2 'B9 + 2 3 B . The result in C 2, is summarized in the left column of Table
0
10-5. In the right-hand column, we connect up the correlating states in , from
Table 10-4. Since we have three triplet states in C2, of Bg, Bg, and Ag symmetry,
2
these must arise from 3Tg. All of the other states from t 2 8 are singlets. No other
1
possible correspondence exists. One can work with the t 28 e. configuration in
a similar manner.
Summarizing the above procedure, we note that we have used the method - 9-
of descent in symmetry as follows:

1. The orbitals in the higher symmetry are correlated with the orbitals in
the lower symmetry.
2. The required number of electrons are added to these lower symmetry
orbitals.
3. The electronic states resulting from the electron configurations in the
lower symmetry are determined.
4. The lower symmetry electronic states, including spin multiplicity, are
correlated with the electronic states of the higher symmetry case. (This
is, in effect, an ascent in symmetry.)
The terms of the d3 strong field configuration are given by taking the direct
product of the terms for d2 with t 2 , and e8 for the added electron.
Next, we have to show how the terms that arise in a strong field are related
to those in a weak field, for there must be a continuous change from one to the
other as a function of the ligand field strength in a series of complexes that a
given metal forms. We can illustrate this with a d 2 ion by employing two principles:
1. A one-to-one correspondence exists between the' states of the same
symmetry and spin multiplicity at the weak field and strong field
extremes.
*lf we label the orbitals 1, 2, and 3, we have I x 1, 1 x 2, 1 x 3, 2 x 2, 2 x 3, and 3 x 3. For

I x 1, 2 x 2, and 3 x 3, we have two electrons in the same orbital.
2. States of the same symmetry and spin degeneracy cannot cross as the
ligand field strength is varied.
Figure 10-5 contains the free ion and corresponding weak field states on
the left and the strong field states on the right. The configurations in an infinitely
strong field are indicated on the far right.
i'A, 'Al
'E e2
3A
- I -
A
I J,
IG TT
FIGURE 10-5 Correlation of 3 T ~

3T
the strong and
2
weak field
states of a d ion. DT
1D
2
'TT2
3T,
Free Weak Strong Infinitely

Ion Field Field Strong
Field
2
If we begin with the infinitely strong field e configuration, we note a 'A
2 shown
state. There is also a 'A, state in t 2 . Only if the connections are made as
3
can we avoid crossing. The same is true for 'E. Since there is only one A 2 state
in the weak field, this connection is straightforward. Proceeding to the other
states from t 2 e and t 2 2 , only the connections shown will lead to non-crossing of
states with the same symmetry and multiplicity.
The results of the evaluation of energies of the various levels in going from
weak field to strong field cases are presented in graphical form in the Tanabe
and Sugano diagrams( 3 ) contained in Appendix C for various d" configurations.
The energies are plotted as E/B versus Dq/B, where B is the Racah parameter.
The quantities E/B and Dq1B are plotted because they enable one to convert the
equation for the energy into a convenient form for plotting. Since states of different
multiplicities are involved, they are a function of the Racah parameter C as well
as B. Therefore, the diagram can be constructed only for a given ratio of C/B.
The lowest term is taken as the zero of energy in these diagrams.
10-5 Double Groups 427
The Orgel diagrams are used to present some of the information in the more
complete Tanabe and Sugano diagrams. Orgel diagrams contain only those terms
that have the same multiplicity as the ground state. Accordingly, they suffice for
the interpretation of the electronic spectra of multiplicity-allowed transitions and
will be employed often in the rest of the chapter (e.g., Fig. 10-9).
10-5 DOUBLE GROUPS
We have previously shown how to use the character tables to find the character
of the representation for which the p and d orbitals form a basis in various
symmetries. In the preceding section, we showed that for any symmetry operation
corresponding to a rotation by an angle a on an orbital or state wave function
having an angular momentum quantum number /, or L, the character x(2) for
which this forms a basis is given by equation (10-9):
sin + I
sin (2)
This equation can also be applied to those states characterized by the total
angular momentum J (where J = L + S) by simply substituting J for /. When
there are an even number of electrons and J is an integer value, the total
representation in any symmetry can be decomposed into the irreducible repre-
sentations of the point group as done in the previous section. However, when J
is half-integral (i.e., S is odd), a rotation by 271 (which is the identity operation)
does not produce the identity for the character:
sin J + I)(a + 27) sin J+ I + J+ 27

x(1 + 27)= _) = -
.o (+ 27r) . a
sin 2 sin 2 +
sin J + 2
sin
It can be shown that rotation by 47r is needed to produce the identity. To
avoid this difficulty, rotation by 27r in this instance is treated as a symmetry
operation that we shall label R. The ordinary rotation group is expanded by
taking the product of R with all existing rotations. The new group is called a
double group. Using equation (10-9), the characters for the rotations can be
worked out. The characters of E and R (i.e., 2 = 0 and 2 = 2r) require evaluation
of the limit of the indeterminate form
sin J+) 27 0
sin r 0
leading to x(O) = 2J + 1 and X(27r) = 2J + 1 for integer J or -(2J + 1) for

half-integer J, as mentioned before. The character tables for the rotation double
groups D4 ' and 0' are given in Appendix B. Those for other groups have been
reported.* Two commonly encountered systems for labeling the irreducible
representations are given. One uses a serially indexed set of F's, and the other
uses primed symbols similar to those we have been employing. The direct products
of representations of double groups can be taken as before and reduced to sums
of irreducible representations.
The above discussion can be clarified by considering some examples. We
need to employ double groups to determine the effects of spin-orbit coupling
when J is half-integral. Since spin-orbit effects arise from coupling of spin and
orbital momenta of electrons, we are concerned with the direct product repre-
sentation of these two effects. As an example, we shall work out the effects of an
octahedral field and spin-orbit coupling on the 4 F free ion state of a d7 ion. As
in the previous section, we can work out the total representation in the 0 point
group and factor it to obtain
x,( = 3) = A 2 + T1 + T2
4
1 2 (4) J = %
-- -------- (4)F - - - - -
/ (4)
/ / / (6)
4
4F (2) ' _ / F (28)
8 -
\ (2) F/
(10)
-- ------- (2) F ///
(4)Fr68 ////
-- -- (2) r' ''7

4\ r 4 (12) _ _(4) r.',",
--- --
----- (2)rg -6
(4) F8'
0
Free Oh Field oh Field , Field AL S> Strong Free
Ion >XL.S zXL-S Oh Field AL-S Ion
(A) (B) (C) (D) (E) (F) (G)
FIGURE 10-6 (A) The gaseous ion, (B) split by a strong O field, (C) followed by
smaller AL S. On the right, (G) free ion, (F) split by large spin-orbit coupling, (E)
followed by a weaker ligand field. Part (D) indicates the correlation of states in the
intermediate region. For convenience, none of the states are shown to cross. States
of different double-group symmetries may cross. States of the same double-group
symmetry will undergo configuration interaction.
* S. Sugano, Y. Tanabe, and H. Kamimura, "Multiplets of Transition Metal Ions in Crystals,"

Academic Press, New York-1970.
10-6 The Jahn-Teller Effect 429
A d7 ion in a weak 0 field leads, as shown in the Tanabe and Sugano diagram,
to a 4 T,, ground state and 'T, and 4Ag excited states. In the 0' double group,
these correspond to T'(F4 ), T2'(Fs), and A2 '(F2 ), respectively. Using S = 32
and substituting S for / in equation (10-9), we generate in the 0' point group
an irreducible representation of 'G(F,), i.e., one of the new irreducible represen-
tations of the double group. Now, we take the direct products of the spin and
orbital parts and decompose them as before, leading to
F 2 x F8 - 8
F4 x 8 = F + F, +2F 8
[5 x 1- = F, + F, + 217
As we see, spin-orbit effects do not split F2, but they split F4 into four states
and F5 into four states. We could have converted L and S to J and employed
equation (10-9) on J values of 9/2, /2, 5/2, and '/2 to obtain the double group
representations. This procedure would have been followed if spin-orbit coupling
were comparable to or greater than the crystal field. Using the approach employed
above, we have assumed a large crystal field and a small spin-orbit perturbation
on it. We can summarize the results with the diagram in Fig. 10-6.
It is important to remember that whenever one is concerned with the effects
of spin-orbit coupling (as we shall often be in subsequent chapters) in a system
with half-integral J values, the double group should be employed.
10-6 THE JAHN-TELLER EFFECT
There is one other effect that influences the electronic structure of a complex,
which we should consider before discussing electronic spectra. Consider a
molecule with an unpaired electron in a doubly degenerate orbital, e.g., an
octahedral Cu(II) system. Note that by distorting the molecule from its most
symmetrical geometry (O) to, say, D4 , it is possible to lower its energy:
eg + bg
{ e.
Oh - D4h
Here the eg set splits into big and aig components. Since two electrons are in the
stabilized a,, orbital and only one is in the destabilized big orbital, the molecule
as a whole is stabilized. It is easy to see how this happens by remembering the
basis of simple electrostatic crystal field theory: an orbital pointing at a ligand
is destabilized. The closer the ligand is, the higher the energy. A tetragonal
elongation (a lengthening of two M-L bonds on the z-axis and shortening of
the other four on the x- and y-axes) destabilizes the d,2_,2 (bag) orbital and
stabilizes the dz2 (aig) orbital. Similarly, a tetragonal compression would raise d 2
and lower dx, -2. Jahn and Teller first pointed out that, for a non-linear molecule,
when such a distortion can occur to lower the energy, it will. We thus expect
that there will be a Jahn-Teller distortion any time we have an orbitally degenerate
(E or T) state and when a proper symmetry vibrational mode exists which enables
the molecule to move from one geometry to the other. One unpaired electron in
a doubly degenerate pair of e orbitals gives rise to an E state and one or two
unpaired electrons in three triply degenerate t orbitals gives rise to a Tstate.
Note that this criterion is very similar to the criterion for spin-orbit coupling.
A simple one-electron picture can be used to predict when orbital angular
momentum contributions are expected and when they are not. To obtain orbital
angular momentum the electron must be in a set of degenerate orbitals that
permit it to move freely from one degenerate orbital to the next and, in so doing,
circulate around an axis. Consider, for example, the d, and d, orbitals of a
metallocene. Degeneracy of this pair permits circulation and angular momentum
about the z-axis. All E and T states will have spin-orbit coupling, except for E
states in the 0 and T point groups. In these latter cases, the E states are composed
of d2 _ , and dz2 so degeneracy does not allow circulation about an axis.
Some people, including Teller, argue that, if a state is split for any reason,
there is no Jahn-Teller effect. Others talk about a Jahn-Teller distortion combining
with other factors that remove degeneracy. This latter approach brings up an
interesting dilemma: When a degenerate state splits, is this due to a Jahn-Teller
distortion, distortion from lower symmetry components in the structure, or
spin-orbit coupling? Since the magnitude of these effects is often comparable (200
to 2000 cm-'), it may only be possible to say that the splitting is due to some
unspecified combination of these effects. One guideline is that Jahn-Teller
distortions are generally larger in E states than in T states, so spin-orbit coupling
is generally the dominant effect in T states.
10-7 MAGNETIC COUPLING IN METAL ION CLUSTERS
In metal clusters containing two or more paramagnetic metal centers, the electron
spins of the individual metal ions S1 , S 2 ... S , interact. The interaction between
a pair of spins i and j is described by the Hamiltonian operator:
f = JS - S (10-10)
where J in energy units is called the exchange coupling constant or exchange

parameter. It is to be distinguished from the J quantum number. With this
Hamiltonian, the exchange parameter J is negative for an antiferromagnetic
interaction, which leads to a pairing of electrons, and is positive for a ferromagnetic
interaction. The literature is confusing in this area, for the above Hamiltonian
has been written with J replaced by 2J, -J or -2J.
Dimeric copper(II) acetate dihydrate is the classic example of this type of
system. The structure of this molecule is shown in Fig. 10-7, and the metal-metal
axis is taken as the z-axis. The copper(II) ions have a d9 configuration. At low
temperatures the compound is found to be diamagnetic, and at elevated
temperatures it is paramagnetic. We can view this system as one in which two
molecular orbitals exist that are largely metal d, -, (with a significant contri-
bution from the bridging acetate). A simplified representation of this part of the
10-7 Magnetic Coupling in Metal Ion Clusters
FIGURE 10-7 The structure of

Cu2(CH3CO2)4 -2 H20. The Cu-Cu
distance is 2.6 A.
molecular orbital diagram of the complex is shown in Fig. 10-8.* (If necessary,
the reader is referred back to the chapter on epr to review the discussion on the d2 2, 1, d 2 2
triplet state and exchange interactions.) With A as the energy separation between
the bonding and antibonding molecular orbitals (Fig. 10-8) and K as the spin
pairing energy, we obtain, when A < K, a ferromagnetic system. When A > K, m.o. 'S
an antiferromagnetic coupled system results. The copper(II) acetate dimer is an FIGURE 10-8 A simplified
example of the latter. In Cu 2 (CH 3 CO 2 )4 -2H 20, the ground state is diamagnetic, representation of the
but the excited triplet state is close by in energy and is thermally populated at interaction of two metal d,2y2
elevated temperatures. The value of J is found 4 ) from magnetic susceptibility orbitals in Cu2. (CH3CO) 4 -2H20
studies (vide infra) to be 284 cm-'. to produce two nondegenerate
In the discussion of the epr spectra of diradicals, new states with S = 0 and levels.
S = 1 were shown to arise (see Fig. 9-22). The general expression for the new
states arising from the exchange coupling of S, and S2 centers is given by:
(SI +S 2),. (Si -S 2)

(10-11)
When S, =S 2 = 2 we obtain S = 0 and S = 1. When S, 2 and S2

= /2 we
obtain S 4, 3, 2 and 1.
We define a total spin quantum number for the system, S, as
S - Si ± S2
Accordingly
S 2 = S12 + S2 2 + 2 5 ,- 2
which upon rearranging defines Si S2 as
- 1[2 _ 12 _ 2
2
] (10-12)
* This is referred to as a super exchange pathway involving the bridging acetates. There is
controversy regarding the amount of a direct exchange contribution, which involves a direct overlap
of the two orbitals on each copper.
mi
--- 1 E= + J + gpH, J+ gpH,
--- 0 E = + {JJ
4J
---------- 0 E=- J 0
S 0 4
Energy Separation
JS- S2 H
9
FIGURE 10-9 Energies of the levels in a dimeric d system.
Recalling that 9 2 0 - S(S + 1)0, we have the following result:
E(S) = JS1 - [S(S + 1) - Si(Si + 1) - S2 (S2 + 1)]0 (10-13)

2
For dimeric copper(II) acetate dihydrate, we have
S 1 = S 2 = 1/2 and S 0 or S=1
For S = 0, we obtain
J [0_3 3] 3
2 4 4 4
whereas for S = 1, we obtain
J 2 3 3] 1
2 4 4- J
By adding gf#,h, to equation (10-10) the Hamiltonian becomes

f = gp3SHz + JS1 - 2 which produces the results shown in Fig. 10-9. Calling
the ground state energy zero, we calculate the energy separation to the excited
states by subtracting the ground state energy.
In a similar fashion the reader is left to show that when a S = '/2 center is
magnetically coupled to an S = '/2 center, equation (10-11) leads to states 4, 3,
2, and 1 with energies from equation (10-14) of 9J, 5J, 2J, and 0. In a magnetic
field, each state will be split into the M, components given by S, (S - 1) ... -S.
The wave functions IS, M,> are expressed as linear combinations of the two
centers IS,, S2, M, Ms12> weighted by their appropriate coefficients C 1 2 -
10-8 Survey of the Electronic Spectra of Oh Complexes 433
10-8 SURVEY OF THE ELECTRONIC SPECTRA OF Oh Applications

COMPLEXES
The electronic spectra of transition metal complexes can be interpreted with the
aid of crystal field theory. In our discussion of 0 , complexes in this section, our
concern will be with systems in which the local symmetry is 0 h, though the
overall molecular symmetry may not be. Throughout the remainder of this
chapter, we shall use the symmetry terms very loosely to describe the type and
arrangements of donor atoms directly bonded to the metal, with the rest of the
ligand atoms being ignored. It should be realized that this assumption is not
always justified. Upon completion of this section, we shall be in a position to
assign and predict the electronic spectrum as well as rationalize the magnitudes
of the d-orbital splittings observed. The treatment here will not be encyclopedic;
selected topics will be covered. The aim is to give an appreciation for a very
powerful tool in coordination chemistry: the utilization of electronic spectra in
the solution of structural problems. More advanced treatments0 containing
5
references to the spectra of many complexes are available.(1,2,4, ,9,i ,12>
The discussion in Chapter 5 of selection rules for electronic transitions should
be reviewed if necessary. Here we shall apply these rules to some transition metal
ion systems. We begin by discussing high spin, octahedral complexes of Mn(II),
a d' case, where there are no spin-allowed d-d transitions. All d-d transitions in
this case are both multiplicity and Laporte forbidden. If it were not for vibronic
coupling and charge transfer transitions, Mn(II) complexes would be colorless.
Hexaquomanganese(II) ion is very pale pink, with all absorption peaks in the
visible region being of very low intensity.
The fact that multiplicity-allowed transitions are usually broad, while
multiplicity-forbidden transitions are usually sharp, aids in making band assign-
ments. Multiplicity-allowed t 2 , - e, transitions lead to an excited state in which
the equilibrium internuclear distance between the metal ion and ligand is larger
than in the ground state. In the course of the electronic transition no change in
distance can occur (Franck-Condon principle), so the electronically excited
molecules are in vibrationally excited states with bond distances corresponding
to the configuration of the ground state. The interaction of an excited state with
solvent molecules not in the primary coordination sphere is variable because
neighboring solvent molecules are various distances away when the excited
molecule is produced. Since the solvent cannot rearrange in the transition time,
a given excited vibrational state in different molecules will undergo interactions
with solvent molecules located at varying distances. Varying solvation energies
produce a range of variable energy, vibrationally excited states and a broad band
results.
In some spin-forbidden transitions, rearrangement occurs in a given level.
For example, in Cr(III) complexes a transition occurs from a ground state
containing three unpaired electrons in t2g to an excited state in which t2, has
two paired electrons and one unpaired electron. In these multiplicity-forbidden
transitions there is often little difference in the equilibrium internuclear distances
of the excited and ground electronic states. Sharp lines result from these transitions
to a low energy vibrational level of an excited state whose potential energy curve
is similar in both shape and in equilibrium internuclear distance to that of the

ground state.
As discussed earlier, there is no center of symmetry in a tetrahedral molecule,
so somewhat more intense absorptions (e = 100 to 1000) than those in octahedral
complexes are often obtained for d-d transitions in Td complexes.
d' and d9 Complexes

The simplest case with which to illustrate the relation between Dq and the color of
a transition metal ion complex resulting from a d-d transition is d' e.g., Ti(III)
in an octahedral field. The ground state of the free ion is described by the term
symbol 2 D and, as indicated earlier, the degenerate d levels are split in the presence
of an octahedral field into a triply degenerate 2T g and doubly degenerate 2
2 E9
set. The splitting is equal to 1ODq. This is represented graphically in Fig. 10-10.
GASEOUS! E
ION
AE
FIGURE 10-10 Splitting of d levels for a

T
d' case, Oh field. 2
uJ I
0 INCREASING Dq
As Dq increases, AE, the energy (hence the frequency) of the transition increases.
The slope of the T2g line is -4Dq and that of Eq is + 6Dq. The value of A (in
units of cm-1) can be obtained directly from the frequency of the absorption
peak. For example, Ti(H 20) 3 has an absorption maximum at about 5000 A
(20,000 cm-'). The A value for water attached to Ti" is about 20,000 cm- ' (Dq
is 2000 cm 1). Since this transition occurs with the absorption of the yellow-green
component of visible light, the color transmitted is purple (blue + red). As the
ligand is changed, Dq varies and the color of the complex changes. The color of
the solution is the complement of the color or colors absorbed, because the
transmitted bands determine the color. Caution should be exercised in inferring
absorption bands from visual observations; e.g., violet and purple are often
confused.
For a d 9 complex in an octahedralfield, the energy level diagram is obtained
by inverting that of the d' complex (see Fig. 10-11). The inversion applies because
the ground state of a d 9 configuration is doubly degenerate [t2geg' can
be t2 g6 (d,2 )2(d, 2 )' or t2g6(d, 2 1 )1(d )2] and the excited state is triply
degenerate [t 2 /5 eg' can be (dx,) 2 (dY) 2 (dxz) 1(eg) 4 or (dy)2(d,,)'(dx) 2(eg) 4 or
(dxy)'(dy) 2 (dz) 2 (eg) 4 ]. Therefore, the transition is 2 E -* T2g. In effect, the
electronic transition causes the motion of a positive hole from the eq level in the
ground state to the t 2 g level in the excited state, and the appropriate energy
- -FIGURE 10-11 Splitting of the d levels

LU in a d' complex, Oh field.
6
Eg(t 1 e.3 )
0 INCREASING Dq
diagram results by inverting that for the electronic transition for a d ' case. In
order to preserve the center of gravity, the slopes of the lines in Fig. 10-11 must
be -6Dq for E and +4Dq for T2 -
The results described above are often summarized by an Orgel diagram as
in Fig. 10-12. For d', the tetrahedral splitting is just the opposite of that for
octahedral splitting, so d' tetrahedral and d' octahedral complexes have similar
Orgel diagrams, as indicated in Fig. 10-12. The splitting of the states as a function
of Dq for octahedral complexes with electron configurations d' and d' and for
tetrahedral complexes with d4 and d' electron configurations is described by the
right half of Fig. 10-12. Only one band arises in the spectra from d-d transitions,
and this is assigned as 2Tg -* 2Eq. The left-hand side of the Orgel diagram applies
to octahedral d 4 and d9 as well as tetrahedral d' and d' complexes. The single
2
d-d transition that occurs is assigned as 2- -2
'' d~d Oh d d60Oh

d1d6T d4d9 T E
FIGURE 10-12 Orgel diagram for high
spin d', d4 , d , and d complexes. (The g
9
---
Z subscript would be added to the
T2 symmetry designation for an Oh complex.)
+-INC Dq 0 INC Dq
d2, d7, d3, and d8 Configurations
The two triplet states for a d 2 gaseous ion were earlier shown to be 3 F and 'P.
Furthermore, we showed that an octahedral field split the 3F terms into the
triplet states 3 Tig, 3T2 g, and 3 A2g; the 3A 2,g state arises from eg2 , the 'Tg states
arise from t 2 g2 , and (though not worked out earlier) another 3 T2g state arises
from t2g t eg'. In understanding the electronic spectra, only the triplet states need
be considered, for the ground state is triplet. A simplified, one-electron orbital
description of these triplet states will be presented. The following degenerate
arrangements are possible for the ground state of the octahedral d 2 complex:
d,1, dxzl, d, 0; d dxz', d,,1; dx,, dxzo, dvz. The ground state is orbitally triply
degenerate and the symbol 3Tig(F) is used to describe this state; the (F) indicates
that the state arose from the gaseous ion F term. In addition to the 3 Tg(F)
ground state, an excited state exists corresponding to the configuration in which
the two electrons are paired in the t2,g level. Transitions to these states are
multiplicity forbidden, but sometimes weak absorption bands assigned to these
transitions are observed.
The triplet excited state t2g'eg will be considered next. If an electron is
excited out of d., or d,, so that the remaining electron is in d,, the excited
electron will encounter less electron-electron repulsion from the electron in d,
if it is placed in d,2 . The d.2 _, orbital is less favorable because of the proximity
2
of an electron in this orbital to the electron remaining in d.. This gives rise to
arrangement (A) in Fig. 10-13. Similarly, if the electron is excited out of d, it
E dm_2 d d2 d d. dz
FIGURE 10-13 Possible N _4__
electron arrangements for R -4 - - - -4-
t29eg G dy dz dy dz dx dyz d, dz dy
Y (A) (B) (C)
will be most stable in d 2 _,2. The remaining electron can be in either d or d ,

giving rise to (B) and (C). This set [Fig. 10-13(A, B, and C)] gives rise to the 3 T,
state, which is orbitally triply degenerate and has a spin multiplicity of three.
The arrangements (d. , de _ ; d, , d, 1; d, , d ) are higher in energy and
also produce an orbitally triply degenerate state, 3T,(P). Other possible arrange-
ments corresponding to t2g'eg' involve reversing one of the electron spins to
produce states with singlet multiplicity. Transitions to these states from the triplet
ground state are multiplicity forbidden. Finally, a two-electron transition produc-
ing the excited state eg2 or d,2, dx2_,2l gives rise to a singly degenerate 3 A 2 g
state. It is instructive to indicate how these states relate to the gaseous ion. As
illustrated in the section on term symbols (Section 10-1), the ground state for
the gaseous ion V(III) (a d 2 ion) is 3 F. The ligand field in the complex removes
the sevenfold orbital degeneracy of this state (i.e., M, = 3, 2, 1, 0, -1, -2, -3)
into two threefold degenerate states, 3 Tg(F) and 3 T2g, and one non-degenerate
state, 3 A2g- This is indicated in the Orgel diagram (Fig. 10-14) for a d 2 , O, complex.
d 2d7 T, d 2d7 O,
d3 d8Oh d 3d 8Td
A2
T,(P) T (P)
T,(F) 3P
T2
FIGURE 10-14 Orgel diagram for high __- 3-T2
spin d2 , d3 , d , and d8 complexes. z T2
LU
<-INC Dq 0 INC Dq-+

TABLE 10-6. Absorption Maxima of Octahedral NI" Complexes

(vmax in cm-1)
3 3 3
Ligand A2g - 3T2g A2g - 3Tig(F) A2, - 3 T1g(P)
H20 8500 15,400 26,000

NH 3 10,750 17,500 28,200
(CH 3)2 SO 7730 12,970 24,040
HC(O)N(CH 3)2 8500 13,605 (14,900) 25,000
CH 3C(O)N(CH 3)2 7575 12,740 (14,285) 23,810
3 3
For zero Dq (i.e., the gaseous ion), only two triplet states, F and P, exist.
3 3 3
As Dq increases, F is split into the T1,(F ), T2 ,, and A 2, states. The degeneracy
3
of the 3P state is not removed by the ligand field, and this state becomes the
triplet 3Tg(P) state in an octahedral complex. The (P) indicates that this state
arises from the gaseous ion 3P state. The energies of these states as a function
of Dq are presented in the Orgel diagrams as well as the Tanabe and Sugano
diagrams ( 3 ) (Appendix D). Use of the Orgel diagrams in predicting spectra and
making assignments will be demonstrated by considering V(III) and Ni(II)
complexes.
For V(III), three transitions involving the states shown in Fig. 10-14 could
3
occur: 3Tl(F) 3 T2 g,g 3Tt(F) -+ Tig(P), and Tg(F) - A2. The transition to
3 3
3
A2,g in V(III) is a two-electron transition. Such transitions are relatively
improbable, and hence have low intensities. This transition has not been observed
experimentally. The spectra obtained for octahedral V(III) complexes consist of
3 3 3
two absorption bands assigned to Tg(f) -- T2 g(F) and Tg(F) - Tg(P). In
3
3
V(H 2 O), + these occur at about 17,000 and 24,000 cm', respectively.
For octahedral nickel(II) complexes, the Orgel diagram (left-hand side of
3 3 3 3T
Fig. 10-14, d') indicates three expected transitions: A2, -+ T2g, 2
and A2, -* Tg(P). (A similar result is obtained from the use of the Tanabe and
3 3
Sugano diagram in Appendix D.) Experimental absorption maxima correspond-

ing to these transitions are summarized in Table 10-6 for octahedral Ni(II)
complexes. (Numbers in parentheses correspond to shoulders on the main band.)
Spectra of the octahedral NH 3 , HC(O)N(CH 3)2 , and CH 3C(O)N(CH 3)2 com-
plexes are given" 4 in Fig. 10-15. These complexes are colored purple, green, and
yellow, respectively.
10-9 CALCULATION OF Dq AND P FOR Oh Ni(II)

COMPLEXES
The graphical information contained in the Orgel diagrams is more accurately

represented by the series of equations that relates the energies of these various
states to the Dq value of the ligand. These energies were derived in Section 10-3.
For Ni(II) in an octahedral field, the energies, E, of the states relative to the
spherical field are given by the following equations.
for 3 Tg: E = -2Dq (10-14a)

for 3 A 2 g: E = -12Dq (10-14b)
3 3
for Tg(F) and Tig(P):
[6Dqp - 16(Dq) 2] + [-6Dq - p]E + E 2 = 0 (10-14c)
Energy (cm')
25,000 12,500 10, 8000
'4 I
FIGURE 10-15 Molar
absorptivity, c, for some 0
nickel(II) complexes in
CH3NO2 solution --- ,
Ni(NH 3)(CIO) 2; 12-
Ni[HC(O)N(CH 3) 2] 6 (C0O4 )2 ; -- ,0 6
Ni[CH3C(O)N(CH 3)2]6 (C1O 4 )2 - //
400 600 800 1000 1200 1400

Wavelength (mp)
where p is the energy of the 3P state. There are two roots to the last equation
corresponding to the energies of the states 3 Tig(F) and 3 T" (P).
From the equations it is seen that the energies of both 3 T,, and 3A 2g are
linear functions of Dq. For any ligand that produces a spin-free octahedral nickel
complex, the difference in energy between the 3 T2 . state and the 3 A 2g state in the
complex is 1ODq. As can be seen from the Orgel or Tanabe and Sugano diagrams,
the lowest energy transition is 3 A g- 3 T2g. Since this transition is a direct measure
of the energy difference of these states, A (or 1ODq) can be equated to the transition
energy, i.e., the frequency of this band (cm-1).
Equation (10-14c) can be solved for the energies of the other states. However,
the above equations have been derived by assuming that the ligands are point
charges or point dipoles and that there is no covalence in the metal-ligand bond.
If this were true, the value for Dq just determined could be substituted into
equation (10-14c), the energy of 3P obtained from the atomic spectrum of the
gaseous ion,( 0 ) and the energy of the other two levels in the complex calculated
from equation (10-14c). The frequencies of the expected spectral transitions are
calculated for one band corresponding to the difference between the energies of
the levels 3 Tg(F) - 3 A2 g and for the other band from the energy difference
3 3
Tig(P) - A 2g. The experimental energies obtained from the spectra are almost
always lower than the values calculated in this way. The deviation is attributed
to covalency in the bonding.
10-9 Calculation of Dq and #l for Oh Ni(lI) Complexes 439
The effect of covalency is to reduce the positive charge on the metal ion, as
a consequence of electron donation by the ligands. With reduced positive charge,
the radial extension of the d orbitals increases; this decreases the electron-electron
repulsions, lowering the energy of the 3 P state. Covalency is foreign to the crystal
field approach and is incorporated into the ligand field approach by providing
an additional parameter, as we shall discuss next.
The difference in energy between the 3P and 3 F states in the complex relative
to that in the gaseous ion is decreased by covalency and, as a result, the gas
phase value cannot be used for p [in equation (10-14c)]; rather, p must be
experimentally evaluated for each complex. Equation (10-14c) can be employed
for this calculation by using the Dq value from the 3A 2 g - 3 T2 g transition and
the experimental energy, AE, for the 3A 2 ->+ Tg(P) transition. The only unknown
remaining in equation (10-14c) is p. The lowering of 3P is a measure of covalency,
among other effects. It is referred to as the nephelauxetic effect and is sometimes
3
expressed by a parameter #l, a percentage lowering of the energy of the P state
3
in the complex compared to the energy of P in the free, gaseous ion."') It is
calculated by using the equation:
#l = [(B - B')/B] x 100
where B is the Racah parameter discussed earlier for the free gaseous ion and
B' is the same parameter for the complex. It should be noted that p of equation
(10-14c) is proportional to B. In the case of nickel(II), the energy of 'P in the
complex can be substituted along with Dq into equation (10-14c) and the other
root calculated. The difference in the energy between this root and the energy of
3
A2 , gives the frequency of the middle band [ 3 A~g -+> 3 Tg(E)]. The agreement of
the calculated and experimental values for this band is good evidence for 0 ,
symmetry. The above discussion will be made clearer by referring to Appendix
3
E, where a sample calculation of Dq, # , and the frequency of the 3A2 g -+ Tiq(E)
transition is presented for Ni[(CH 3)2 SO]6 (Cl1 4 )2 .
Most often the quantity # is used instead of #l , where #l is defined as:
B' (10-15)
B
The two quantities are easily related if equation (10-13) is rewritten as

S-(1 - #) x 100.
With many other ions the spectral data cannot be solved easily for Dq and
# because of complications introduced by spin-orbit coupling. The consequences
of this effect on a d' ion are illustrated in Fig. 10-16. The triply degenerate T2g
state is split by spin-orbit (s.o.) coupling as indicated in Fig. 10-16(B). Coupling
lowers the energy of the ground state and the extent of lowering depends upon
the magnitude of the coupling. When the ground state is lowered by spin-orbit
coupling, the energies of all the bands in the spectrum have contributions from
this lowering, Ae . When the contribution to the total energy from A,. cannot
be determined, the evaluation of A and # is not very accurate. Spin-orbit coupling
in an excited state is not as serious a problem because transitions to both of the
split levels often occur and the energies can be averaged. When the ground state
is split, only the lower level is populated. Thus accurate values for Dq and #
without corrections for spin-orbit coupling can be obtained only for ions in which
FIGURE 10-16 Contribution to E

A from spin-orbit coupling. (A)
N pen
d-Level splitting with noE
spin-orbit coupling. (B) Splitting R
of T2g level by spin-orbit coupling. - -A
JT2
A
(A) (B)
the ground state is A or E (e.g., Ni(II)). A further complication is introduced by

the effect that Jahn-Teller distortions have on the energies of the levels.
Ni(II), Mn(II) (weak field), Co(III) (strong field), and Cr(III) form many
octahedral complexes whose spectra permit accurate calculation of Dq and #
without significant complications from spin-orbit coupling or Jahn-Teller distor-
tions. Ti(III) has only minor contributions from these complicating effects. In the
case of tetrahedral complexes, the magnitude of the splitting by spin-orbit
interactions more nearly approaches that of crystal field splitting (Dq', the splitting
in a tetrahedral field, is about 49 Dq). As a result, spin-orbit coupling makes
appreciable contributions to the energies of the observed bands. A procedure has
been described" 4 ) that permits evaluation of Dq and # for tetrahedral Co(III). A
sample calculation is contained in Appendix E.
Both a and 71 bonding of the ligand with the metal ion contribute to the
quantity Dq. When 7[ bonding occurs, the metal ion t~g orbitals will be involved,
for they have the proper directional and symmetry properties. If 7T bonding occurs
with empty ligand orbitals (e.g., d in Et 2 S or p in CN-), Dq will be larger than
in the absence of this effect. If 7z bonding occurs between filled ligand orbitals
and filled t2,g orbitals (as in the case with the ligand OH~ and Co(III)), the net
result of this interaction is antibonding and Dq is decreased. These effects are
illustrated with the aid of Fig. 10-17. In Fig. 10-17(A), the d electrons in t,
interact with the empty ligand orbitals, lowering the energy of t2g and raising
the energy of the mainly ligand 7r orbitals in the complex. An empty nT*orbital
of the ligand could be involved in this type of interaction. Since t2,g is lowered
in energy and eg is not affected (the eg orbitals point toward a electron pairs on
the ligands), Dq will increase. In the second case [Fig. 10-17(B)], filled ligand 71
orbitals interact with higher energy filled metal d orbitals, raising the energy of
t 2, and lowering Dq.
It is informative to relate the energies of the observed d-d transitions to the
energy levels associated with the molecular orbital description of octahedral
complexes. The scheme for an 0 , complex is illustrated in Fig. 10-18 (which
neglects 7nbonding). The difference between T2 , and E,* is 1ODq. As metal-ligand
a bond strength increases, E9 is lowered, E9* is raised by the same amount, and
Dq increases. If T2g metal electrons form 7r bonds with empty p or d orbitals of
the ligand, the energy of the T2g level in the complex is lowered and Dq is increased.
Electron-electron repulsions of the T2g electrons and the metal nonbonding
electrons raise the energy of the T2g set and decrease A. The above ideas have
been employed in the interpretation of the spectra of some transition metal
acetylacetonates. (15,16,17)
10-9 Calculation of Dq and # for Oh Ni(ll) Complexes 441
77
/eg 77
e Ag
FIGURE 10-17 Effect of it bonding

on the energy of a t2, orbital and on
Dq. (A) Filled metal orbitals, empty
ligand orbitals. (B) Filled metal
orbitals, filled ligand orbitals.
t2 9
(B)
Complex, Complex, Ligand
no iT including 7T 7 orbital
bonding bonding
4ptl, ---
FIGURE 10-18 Molecular

4s a1g orbital description of an
octahedral complex (it
bonding effects and
electrons are not
included).
The magnitude of Dq is determined by many factors: interactions from an

electrostatic perturbation, the metal-ligand o- bond, the metal to ligand it-bond,
the ligand to metal it bond, and metal electron-ligand electron repulsions.
Additional references(9,10,18) on the material presented in this section are available.
Much useful information regarding the metal ion-ligand interaction can be
obtained from an evaluation of Dq and #l. For a series of amides of the type
R2
RC(O)N it was found that whenever R, and R 2 are both alkyl groups, lower
~R3
values for Dq and # result for the six-coordinate nickel complexes than when
either R, or both R 2 and R3 are hydrogens. This is not in agreement with the
observation that toward phenol and iodine the donor strengths of these amides
are found to increase with the number of alkyl groups. It was proposed that a
steric effect exists between neighboring coordinated amide molecules" in the
metal complexes. A study of the nickel(II) complexes of some primary alkyl amines
indicated that even though water replaces the amines in the complexes, the amines
interact more strongly with nickel than does water, and almost as strongly as
ammonia.(19 ) A large Dq is also reported for the nickel complex of
ethyleneimine.t2 ) These results are interpreted and an explanation involving
solvation energies is proposed for the instability of the alkylamine complexes in
water.(19)
The magnitude of 1ODq for various metal ions generally varies in the following
order:
Mn(II) < Ni(II) < Co(II) < Fe(II) < V(II) < Fe(III) < Cr(III) < V(III)
< Co(III) < Mn(IV) < Mo(III) < Rh(III) < Pd(IV) < Ir(HI) < Re(IV)
< Pt(IV)
Representative ligands give rise to the following order, referred to as the

spectrochemical series:
2
I- < Br- < -SCN < F- < urea < OH - < CH 3 CO 2 < C2 0 - <
H 2 0 < -NCS~ < glycine < CHN ~ NH 3 < ethylenediamine <
SO 32 - < o-phenanthroline < NO 2 < CN-
TABLE 10-7. Empirical Parameters for Predicting 1ODq and Bwith Equations
(10-16a) and (10-16b)a
3
Ligands f h Metal Ions g(10 cm-1) k
6F- 0.9 0.8 V(II) 12.3 0.08

6H 20 1.00 1.0 Cr(III) 17.4 0.21
6 urea 0.91 1.2 Mn(II) 8.0 0.07
6NH 3 1.25 1.4 Mn(IV) 23 0.5
3 en 1.28 1.5 Fe(III) 14.0 0.24
3ox 2 - 0.98 1.5 Co(I1I) 19.0 0.35
6Cl- 0.80 2.0 Ni(1I) 8.9 0.12
6CN- 1.7 2.0 Mo(III) 24 0.15
6Br 0.76 2.3 Rh(III) 27 0.30
3dtp 0.86 2.8 Re(IV) 35 0.2
C5 HN 1.25 - Ir(III) 32 0.3
Pt(IV) 36 0.5
" From C. K. Jorgensen, "Absorption Spectra and Chemical Bonding in Complexes," Pergamon
10-10 Effect of Distortions on the d-Orbital Energy Levels 443
Jorgensen(4ab) has reported a remarkable set of parameters that enable one
to predict 1ODq and # for various transition metal ion complexes. When the
empirical parameters given in Table 10-7 are substituted into equations (10-16a)
and (10-16b), the values of 1ODq and B for the complex result.
1ODq =fg (cm-' x 1-3)
B = B 0 (1 - hk)
B0 is the free ion interelectronic repulsion parameter.
10-10 EFFECT OF DISTORTIONS ON THE d-ORBITAL

ENERGY LEVELS
Since octahedral, square planar, and tetrahedral crystal fields cause different
splittings of the five d orbitals, geometry will have a pronounced effect upon the
d -> d transitions in a metal ion complex. Spectral data for these transitions should
provide information about the structure of complexes. Our initial concern will
be with how structure affects the energies of the various states in a metal ion.
This information will then be applied to determine the structures of various
complexes.
The structures of six-coordinate complexes can be classified as cubic, axial,
or rhombic, if the equivalences of the ligands along the x, y, and z axes are
represented by x = y = z, x = y : z, or x 5 y 5 z, respectively. Tetragonal and
trigonal distortions (i.e., an elongation or compression along the threefold axis)
are commom axial examples. The splitting of the states for a d' case is represented
in Fig. 10-19. Since d-d electron repulsions are not present in the d' case, the
B1
E E9
N
E 2
D d22 FIGURE 10-19 Orbital splitting for d'
R complexes in octahedral and tetragonal
G T2 g fields.
y
-E
Gaseous 0
h Tetragonal
ion
states can be correlated with the d orbitals as indicated. For the splitting in a
tetragonal complex, trans-TiA 4 B2 3 , to arise as indicated in Fig. 10-18, ligand
A must occupy a higher position in the spectrochemical series than B. Metal
electron-ligand electron repulsions are less for states consisting of electrons in
orbitals directed toward the ligands, B, located on the z-axis. As a result, d.2 is
lower in energy than d g,2and dx, and d,, are lower than d,,. More bands will
I
be observed in the spectrum of the tetragonal complex than in the spectrum of

the octahedral complex.
E 1
N ~~9
FIGURE 10-20 Splitting of the E 'E

various states for cobalt(Ill) in R -F -
octahedral, tetragonal, and G 11 AE
rhombic fields. Y I --------
A, 'A19
Oh Tetragonal Rhombic
The energies and splittings of the various states for a spin-paired Co(III)
complex are indicated in Fig. 10-20. Since this is an ion with more than one d
electron, we must concern ourselves with states and not orbitals. The 'T, excited
state of the octahedral complex splits into 'A2g and 'E, states in a tetragonal
field, while 'T2g splits into 'E, and 'B 2,. The splitting that occurs in a rhombic
field is also indicated. Hydrated tris(glycinato)cobalt(III) exists as a violet 0 isomer
and a red # isomer. One isomer must be cubic (where x = y = z), and the other
isomer must be rhombic. The spectrum(18) of the # isomer consists of two bands.
The a isomer also gives rise to two bands, one of which is asymmetric and must
consist of two or more absorption bands that are not resolved. Therefore, the a
isomer must be the rhombic isomer and the #l isomer the cubic isomer.
In the cis- and trans-CoA4B 2* complexes, the difference in energy
between the 'A2g and 'Eg states (AE, in Fig. 10-20) is usually about twice as
21 22
large in trans complexes as the corresponding transition in cis complexes )
[AE,(cis) = - C(AA - A.) and AE,(trans) = 2C(AA - AB), where C is a constant

usually less than one, AA and AB represent the crystal field splittings of ligands
A and B (i.e., their positions in the spectrochemical series), and the minus sign
accounts for the fact that the energies of 'E, and 'A 2. are interchanged in cis
and trans complexes]. Usually, when AA and A, differ appreciably, the 'A 2g and
1E. splitting gives rise to a doublet for the 'A,, -- 'Tig peak in the trans compound,
while this band is simply broadened in the cis compound.( 22 > It also found that
cis isomers often have larger molar absorptivity values for d -+ d transitions than
trans isomers. Typical spectra(23 ) are illustrated in Fig. 10-21. When A, and AB
are similar, this criterion cannot be employed. The ultraviolet charge transfer
band can also be used to distinguish between cis- and trans-cobalt(III)complexes
when both isomers are available. The frequency of the band is usually higher in
the cis than the trans compound. The benzoylacetonates of Co(III) and Cr(III)
are examples in which A, and AB are nearly equal and the meridional (trans)
2
complex has a larger E value than the facial (cis) complex. 4)
If certain assumptions are made concerning the metal-ligand distance, the
dipole moment of the ligand, and the effective charge of the nickel nucleus, it is
possible to calculate12 ) the change in energy that occurs for the various states
as the ligand arrangement is varied from 0 to D4 . This corresponds to
lengthening the metal-ligand distance along d,2 to infinity. The change in energy
for the various levels as a function of the distortion is illustrated in Fig. 10-22.
The percentages on the abscissa indicate progressive weakening of two trans
10-10 Effect of Distortions on the d-Orbital Energy Levels 445
Co(en)3
100 0.005 M Co(en) 3 (CY1
4)3
(A)
e 0
cm FIGURE 10-21 Spectra of

100- (A) octahedral
(B) E cis-Co(en) 2F'2 trisethylenediamine Co(Ill),
(B) the cis difluoro
compound, and (C) the trans
cm difluoro compound. All
spectra are taken in water.
(C) e trans-Co(en) 2F
n
10,000 15,000 20,000 25,000 30,000 35,000
cm-
metal-ligand bonds (e.g., 100% weakening represents a square planar complex).

It is thus easy to see why the spectrum of a square planar or tetragonally distorted
nickel complex should differ from that of a regular octahedral complex.
As indicated in Fig. 10-22, there will be a continuous change in spectral
properties as the amount of tetragonal distortion increases. Eventually, for highly
distorted tetragonal complexes, the spectra will resemble those of square planar
complexes. With large distortion, the multiplicity of the lowest energy state for
Ni(II) becomes singlet and a diamagnetic complex results. The diamagnetic
Eg
3
A2 _ T1g P)
Eg 3Tig (F) FIGURE 10-22 Effect of tetragonal

B 2g
distortion on the energy levels of
9
(AF) -- - -- 5(Eg)
--- nickel(II). [From C. Furlani and G.
l' 9 ) Sartori, J. Inorg. Nucl. Chem., 8,
126 (1958).]
A29
819
BA2g
D4th
.
I
I i I I
100 80 50 20 0
Percent distortion from Oh symmetry
tetragonal or square planar complexes have high intensity absorption bands

(P= 100 to 350) with maxima in the 14,000 to 18,000 cm ' region. The spectra
26 2
may contain one, two, or three peaks, , 7) and band assignments are often
difficult. However, by using spectral and magnetic data, square planar or highly
distorted tetragonal nickel(II) complexes can be easily distinguished from nearly
octahedral or tetrahedral complexes.
Just as distortion from 0, and D4, changes the energies and properties of
the various levels, so does distortion from D4, (planar) to D2 to T (Fig. 10-23).
Eg
A2g 3
Ti
T2g)
----
E
FIGURE 10-23 Energy N 3A
levels for Td, D4h, and D2 d E
complexes. {From C. Furlani R E 3
and G. Sartori, J. inorg. Nucl. G 2g 3
Chem., 8, 126 (1958).] y
3
T9
B 29
D4 D 2d Td
25 28
Figure 10-24 contains spectra for Oh, D 4 h, D 2 d, and Td complexes.( , 30) The
2
transitions of the Td nickel complex NiCl 4 have large E values because the
complex does not have a center of symmetry. In this case, the d and p orbitals
can mix in a molecular orbital description. The p orbital contribution of the
ground and excited states gives some allowed d -* p character to the transition,
and the intensity increases. Mixing in non-centrosymmetric ligand molecular
orbitals will also enhance the intensity. Accordingly, E/5 is plotted in Fig. 10-24.
As can be seen, different spectra are obtained for different structures. The spectrum
for a Td complex is expected (see Fig. 10-23) to contain three bands v1, v2 , and
3 3
v3 corresponding to the three-spin-allowed transitions: T,(F) -+ T2 , v,;
3 3 3
3TI(F) -+ A 2 , v2 ; and TI(F) - T1(P), v3 (see Fig. 10-24). The v, band occurs in
the range between 3000 and 5000 cm-' and is often masked by absorption by
either the organic part of the molecule or the solvent. It has been observed for
Ni(II) in silicate glasses and in NiCl42 . The v2 band occurs in the 6500 to 10,000
cm region and has small molar absorptivity (e = 15 to 50). The v3 band is
found in the visible region (12,000 to 17,000 cm-) and shows moderate absorption
(e = 100 to 200).
It is proposed1 28 ) that the complex Ni[OP(C6 H,),] 4 (ClO4 )2 has a D 2d
configuration. Absorption peaks occur in the spectrum at 24,300, 14,800, and
13,100 cm-' with e values of approximately 24, 8, and 9, respectively.
S35- FIGURE 10-24 Electronic

.5 absorption spectra of some
\nickel complexes. (A)
.5 Ni(03 P0) 4(Cl0 4 ) 2 in CH 3NO
)~0A/ (D2d); (B)Ni[(CH 3 )2 S016 (Gl0 4) 2
/in (CH 3 )2 S0 (h); (C) NCI2
02 //- \ \
10 / in 0H 3N0 2(Td); (D), Ni(lI)
104 -- A ~ (dimethylglyoxinate) 2 in
I~ 7
-CHC
3 (DW. Curve Cisa plot
400 500 600 700 800

Wavelength (my)
10-11 STRUCTURAL EVIDENCE FROM THE ELECTRONIC

SPECTRUM
The electronic spectrum can often provide quick and reliable information about
the ligand arrangement in transition metal ion complexes. Tetrahedral complexes
are often readily distinguished from six-coordinate ones on the basis of the
intensity of the bands. The spectra of nickel(II) and cobalt(II) are particularly
informative. The complex Ni{OP[N(CH,)2 ]3 4 C1( could have tetrahedral, DN2d'
square planar, tetragonal, or other distorted octahedral geometries. The similarity
of the electronic spectrum of this complex to that of NiC( 4 [see Fig. 10-24(N] C)
implied(30 that this was the first cationic, tetrahedral nickel(II) complex ever i
prepared. Further confirmation of the structure comes from the similarity of the
x-ray powder diffraction patterns of the nickel(II) and zinc(Il) complexes. The
latter, with a 3d' 0 configuration, is expected to be tetrahedral. The Orgel or
Tana(e-Sugano diagram for a Td, dd complex is the same as that for octahedral
cobalt(II) (OhCdH)with a low Dq value. Accordingly, the high energy visible band
is assigned to 'T 1(F) --+ 'T 1(P) and the low energy band to 'T 1(F) -- 4 A, The *
commonly observed splitting of the visible band is attributed to spin-orbit

coupling, which lifts the degeneracy of 4 Tu(P) state. It is recommended that any
discussion of band assignments in this section be accompanied by reference to
the Tanabe and Sugano (or Orgel) diagrams.
In another complex, Ni(N it was shown 3 1 that the electronic
H))2}C,
spectrum of the nickel is that of a six-coordinate complex. Some of the nitrate
groups must be bidentate. The color of a transition metal ion complex is often
a very poor indicator of structure. Octahedral nickel(II) complexes usually have
three absorption bands in the regions from 8000 to 13,000 cm ',from 15,000 to
19,000 cm, and from 25,000 to 29,000 cm'. The exact position will depend
upon the quantities A and f. The molar absorptivities of these bands are generally
below 20. The ligand field fit of the calculated and experimental frequencies of
the middle peak has been proposed as confirmatory evidence for the existence
0
of an , complex.
Spin-free tetragonal nickel(II) complexes, in which the two ligands occupying
either cis or trans positions have Dq values that are similar to the other four,
will give spectra that will be very much like those of the 0 complexes. In general,
molar absorptivities will be higher for tetragonal than for octahedral complexes.
A rule of average environment relates the band maxima in these slightly distorted
tetragonal complexes to the Dq values of the ligands. The band position is
determined by a Dq value that is an average of all the surrounding ligands.(4,3 2 )
Nickel(II) forms a large number of five-coordinate complexes. (3 3 Geometries
based on both the trigonal bipyramid and the tetragonal (square) pyramid are
35
known. Many of the complexes are distorted significantly from this geometry.( )
3
The electronic spectra have been analyzed in detail by Ciampolini,( s> and the
interested reader is referred to his account. It is often difficult to detect the
difference between tetrahedral and certain five-coordinate geometries on the basis
of their electronic spectra.
The electronic spectra of cobalt(II) complexes can often provide reliable
structural information. Most six-coordinate complexes are high spin. The Orgel
diagram is given in Fig. 10-14. The ground state is 4 Tg and a substantial amount
of spin-orbit coupling is expected. Three transitions are predicted, 4 T ,(F) -+4T",
4
Tq(F) -4 A2g, and 4 T,,(F) , 4 T(P). The 4 T,g(F) - 4 A 2q transition is a two-
electron transition and is not observed. The electronic spectrum of octahedral
Co(H 2 O) 2 + and tetrahedral CoCl 4 2 - are shown in Fig. 10-25. The band in the
octahedral complex at ~20,000 cm-1 is assigned as the 4 Tg(F) -+ T1,(P)
4
4
transition. The shoulder results because spin-orbit coupling in the excited Tg(P)
state causes the degeneracy to be lifted. The other absorption band, at 8350 cm 1
is assigned to 4 Tg(F) - 4T.
Tetrahedral complexes of Co(II) have an energy level diagram like that for
Cr(III). The complexes will always be high spin (see the Tanabe and Sugano
diagrams in Appendix D). The absorption band at ~ 15,000 cm- 'is assigned to
4
4A2 + T,(P). The fine structure is attributed to spin-orbit coupling of the T
state. The existence of spin-orbit coupling also allows some quartet -+ doublet
spin transitions to occur. The other band shown is assigned to 4 A 2 -+ 4 T1 (F).The
expected 4A 2 - 4 T 2 transition is predicted to occur at 3000 to 4500 cm-1; this
is outside the range of most u.v.-visible instruments and is often overlapped by
ligand infrared bands. Several five-coordinate complexes of cobalt(II) have been
prepared and their spectra reported and interpreted.
Copper(II) complexes( 3 sb) take on a wide range of geometries, often with
low symmetry. One generally finds, for most geometries, one very broad band
with a maximum around 15,000 + 5,000 cm-', which is thought to contain all
of the expected transitions. It is possible, however, that the highest-energy
transition occurs farther out in the ultraviolet region under charge transfer bands.
Thus, the electronic spectrum of copper(II) is of little value in structure assignment.
The band position can be correlated roughly with the ligand field strength of the
bonding groups.
Spectral data are available to enable similar conclusions to be drawn
concerning the structures of other transition metal ions. The salient differences
between spectra for various structures have been summarized."') Infrared and
magnetic data should be used in conjunction with electronic spectral data to aid
Co2 (d7 )
cm-1
(A)
600o FIGURE 10-25 The

electronic spectra of (A)
Co(H20)6 2 * [0.021 M Co(BF 4)2
500- in H20]; (B)CoC142 [0.001 M
(CoC12-I (dl) CoC12 in 10 M HC]. The
dotted line in (A) gives the
400- resolution of the two bands
contributing to the observed
spectrum.
300 -
200-
100-
5,000 10,000 15,000 20,000 25,000

cm-1
(B)
in assigning structures to complexes.136 >Use of magnetic data will be described

in the next chapter.
The above examples are only a few of the very many cases that indicate the
utility of near infrared, visible, and ultraviolet spectroscopy in providing infor-
mation about the structures of complexes. The number of bands, their frequencies,
and their molar absorptivities should all be considered. Solution spectra should
be checked against the solid-state spectra (reflectance or mulls) to be sure that
changes in structure are not occurring in solution. These changes could involve
ligand rearrangement, ligand replacement by solvent, or expansion of the
coordination number by solvation.
Other kinds of structural applications of visible spectroscopy have been
reported. The Dq values for the nitrite ion are different for the nitro (-NO 2)
and nitrito (-ONO) isomers. As a result of the difference in average Dq,
[(Co(NH 3)sONO] 2 + is red, while [Co(NH 3),NO2]2+ is yellow. Limitations of
this kind of application have been reported.(")
The use of electronic spectra to provide structural information is nicely
illustrated in a study of the electronic structure of the vanadyl ion.(38 ) Spectra
of the vanadyl ion, V0 2 ', are interpreted to indicate that there is considerable
oxygen-to-metal 7r bonding in the V-O bond. The similarity in the charge transfer
2
spectra of solids known by x-ray analysis to contain the V0 + group and of
solutions is presented as evidence that aqueous solutions contain the species
2
VO(H 2 O),2+ and not V(H 2 O)6 4* or V(H 2O) 4 (OH)22+. Protonation of VO +
would have a pronounced effect on the charge transfer spectrum. It is proposed
that the oxygen is not protonated because its basicity is weakened by 7Ebonding
2
with vanadium. A complete molecular orbital scheme for VO(H 20) + is
3
presented,< s" and assignments are made for the spectrum of VOSO 4 -5H20 in
aqueous solution. Similar studies on other oxy-cations provide evidence for
39
considerable metal-oxygen 7r bonding( ) and aid in elucidating the electronic
structures of these species.
The electronic spectrum has been particularly valuable in determining the
coordination number and ligand arrangement in metallo-enzymes. When zinc(II)
is replaced by cobalt(II) in carbonic anhydrase, the electronic spectrum indicates
that the metal ion is in a distorted tetrahedral site.14 01 In such applications, one
must be particularly careful to ascertain that the structure of the enzyme has not
been changed by metal substitution. If the enzyme is still active, one can have
some confidence that the structure is the same. In this example, subsequent x-ray
structure determination confirmed the distorted tetrahedral ligand arrangement
around zinc(II). The band position in the visible spectrum of a copper(II) protein,
erythrocuprein, was interpreted(") as indicative of coordination by at least four
nitrogen donor ligands.
The blue copper(II) protein stellacyanin has been converted to a cobalt(II)
derivative.("') Copper(II) and cobalt(II) were shown to compete for the same site
in the protein. Since cobalt(II) spectra are more readily interpreted than those
of copper(II), the authors were able to conclude that the cobalt was in either a
distorted tetrahedral or a five-coordinate environment. A strong charge transfer
band indicated the existence of a Co-SR linkage. All of the bands in the native
copper protein were assigned by analogy. The existence of porphyrin complexes
in an enzyme system can be detected by the characteristic Soret band around
25,000 cm-'. This is a ligand-based 7r -+n* charge transfer type of transition,
discussed in Chapter 5. Two other lower intensity bands are also found in the
electronic spectra of these complexes. The existence of these bands and their shifts
upon placing substituents on the rings are understood in terms of results from
molecular orbital calculations.1 2 1 The positions of these bands have been
employed to classify a whole host of cytochromes.
Bonding 10-12 - AND 7 BONDING PARAMETERS FROM THE

Parameters SPECTRA OF TETRAGONAL COMPLEXES
from Spectra As mentioned earlier, when the symmetry of a complex is lowered from a local
0, or T environment additional bands appear in the spectrum. This is illustrated
2
in Fig. 10-25, where the spectra of Co(NH 3) 6 + and Co(NH 3)sCl + are compared.
3
As indicated by the Tanabe and Sugano diagram (Appendix D), for a strong field
d' complex the spin-allowed transitions are 'A1, -+ 'T1 g and 'A 1 , -+ 'T 2, and the
bands are assigned as indicated. When one of the ammonia molecules in the
10-12 a and 7r Bonding Parameters from the Spectra of Tetragonal Complexes 451
complex is replaced by a different group, which we arbitrarily locate on d.2, that

group interacts differently with cobalt than does ammonia. If this group were
chloride, the a interaction would be weaker and the a interaction greater. The
relative energies of d,2, d.,, and d,, will be affected differently in Co(NH 3)5 Cl 2 +
and Co(NH 3)6 3 +; the stronger 71 interaction of Cl- raises the energy of d., and
d,, and a weaker a interaction from Cl- lowers that of dZ2. The degeneracy of
the triplet state is removed and 'T 1 is split into 'E and 'A states, producing the
spectrum indicated in Fig. 10-26(B).
1 T,
(A)
FIGURE 10-26 Spectra of (A)
Co(NH3) 63+ and (B) Co(NH 3)sCl 2 +
LU
'T2
(B)
Frequency, v
The splitting of the 'T 2 band is predicted by theory( 2 2 > to be too small to
be observed. With the additional bands, there is need for additional parameters
(besides 1ODq and B) to describe the spectrum. The calculation of values for these
parameters depends very much upon the assignment of the observed electronic
transitions.
Several lower symmetry arrangements are treated by Gerloch and Slade.(3)
Here, we shall briefly consider tetragonally distorted six-coordinate complexes.
The total potential is now considered as
V(D 4 6) = Vc: + Kewr (10-17)
When the matrix elements are evaluated, we obtain quantities that we earlier
labeled Dq, plus two other radial quantities that arise from V, which we shall
label Ds and Dt. Ds involves r 2 terms and Dt involves r 4 terms, as shown in
equations (10-18) and (10-19):
2 2- 1
Ds = Ze2 r ( 3 1
b3 (10-18)
2 (1 1
Dt = Ze 2 r4 (10-19)
7 a(xy) ()
Here a and b refer to the different metal-ligand distances (a refers to xy-ligands

and b refers to z-ligands). Dt, being an r- function, can be related to Dq. (Recall
that Dq is a function of r~4.) It is a measure of the difference in Dq between the
axial and equatorial sites, as shown in equation (10-20):
4
Dt - [Dq(xy) - Dq(z)] (10-20)
7
It is to be emphasized that neither Dq(xy) nor Dq(z) is the same as that for the
ligand in an octahedral complex.(43 )
Assuming that the symmetry of the complex is such that the metal eg orbitals
are only a-antibonding and that the metal t 2 g orbitals are only 7E-antibonding,
McClure 2 2 ) has reported parameters ba and 67n, defined as:
12 15
60- = c - UX, = - 8 Ds - Dt (10-21)
8 8
3 5
on = 7 - , =- 3Ds + 2 Dt (10-22)
The a and 7r parameters reportedly indicate the relative a and 71 antibonding

properties of the ligands. Values for various ligands have been reported and
interpreted.( 4 3 -4s) The following order of a bond interaction with the metal ion
results from spectral studies on these complexes:
NH 3 >H 2 0>F- >Cl >Br- >I-
The order of 7r repulsion is:
I > Br > Cl > F- > NH 3
10-13 THE ANGULAR OVERLAP MODEL
An alternative approach (with several advantages) to the parameterization of the

spectra of transition metal complexes is the angular overlap model.(3,4'> This
model has its origins in a crude molecular orbital treatment of the energies of
the transition metal compounds. A simple case, which we will consider first, is
that of a monocoordinated complex ML:
M-L
If the metal is a transition metal, we are most interested, from a spectroscopic

point of view, in the energies of the d orbitals in the complex. The five d orbitals
in the C., symmetry of the complex span the a, 7r, and 3 representations; i.e.,
d(z 2 ) is a, d(xz) and d(yz) are it, and d(xy) and d(x2 - y 2 ) are 6. Considering, for
instance, the a interaction, we can write the secular equations:
HMa - E HMLa SML"E 0 (10-23)

HML6 - SML'E HLa - E
where HM" and HL" are the energies of the appropriate metal and ligand orbitals
respectively; HMLO describes the exchange integral between the metal and ligand
orbitals, and SMLO is the overlap integral. In general, it is found that H " HL'
and that the diatomic overlap integral SMLO is small, so that one of the roots, say
E1 , will be quite close in energy to H,", and the other, E2 , will be quite close to
HLa. Invoking this assumption enables us to write two approximate determinants:
H,- - E, HMLO - SML'HM" 0

HMLO - SML-HMa HL" - H"
Hm" - HLa HMLO - SML'HLa 0 (10-24)

HMLO - SML'HLO HL" - E
The values of the roots can be written explicitly as
2 2
E = Hm" + (HML ) - 2HMLHMSML+ (SMLHm) (10-25)
HM" - HLa
2
E2 = HL - (HML )
2
- 2H HLSM MLHL) (10-26)
HMO HLa
Using the Wolfsberg-Helmholz approximation for HMLO in the form
HML SML"(HMa + HLa) (10-27)
equations (10-25) and (10-26) can be rewritten as
2 2
El - Hm" = E,* = (Hma + HL' (SMLa)
HM - HL(
(HM" + H~a)2
E2-HLO=E - - , H (SML )2 (10-29)
Ha - H
HM~ HL M
E,* is positive, and therefore represents the destabilizing effect on the metal
orbital energies, whereas E. is negative and represents the stabilizing effect on
the ligand orbitals. As is shown by equation (10-28), the destabilizing effect on
a particular metal orbital is proportional to (SMLa) 2 . This effect has to be small,
when Hm" - HLa is large and LO is small. E,* might in principle be calculated
using (10-28), but in practice is more profitable to express it parametrically.
it
For this purpose, the overlap integral SMLO can be factored into a radial and an
angular product:
SMLO SMLFO' (10-30)
where Sm is the integral of the radial functions of the metal and of the ligand
orbitals, and F. refers to the angular part. SML depends on the particular metal
and ligand orbitals considered and on the metal-ligand distance, whereas F,.
depends only on the geometrical dispositions of the metal and the ligand. Once
the geometry is known, Fd can be easily calculated. We can demonstrate this
factoring with a simple example of a ligand L overlapping a p, orbital on a metal,
M, in an M-L fragment. In Fig. 10-27(A), the ligand is located on the z-axis
and the M-L bond is coincident with this axis. If we define FfP (sigma ligand
overlap with a p orbital) as one, the overlap is given by the magnitude of SML.
When the M-L bond is the same length but not coincident with the z-axis, the
PZ net overlap will be decreased. The radial part SML stays the same, so the decrease
is accomplished by decreasing F, the angular part. In Fig. 10-27(B), the M-L
bond is along the x-axis, the ligand a orbital is orthogonal to p. (overlap is zero)
and, since the radial part doesn't change, this is accomplished by having the
angular term, F/, become zero.
(a) Substituting equation (10-30) into (10-28), one gets
PZ
(Hm" + HL, 2 d2
E,*=H HL, SML2(Fq) (10-31)
Now letting
(b)
FIGURE 10-27 Orbital (Hma + HLa)2

overlap with a p, orbital in e
e, =
H2
ML2( g)
H(
M-L fragments.
equation (10-31) becomes
E,* = e,(Fd) 2 (10-33)
Equation (10-33) shows that the a-antibonding effect on a particular d orbital

can be expressed by a parameter, e, and a number (Fqd), which can be obtained
from standard tables (vide infra).
In the case of it-bonding, the ligand can interact either through ligand low
energy filled orbitals, so that the same conditions hold as for the a case, or
through ligand high energy empty orbitals, so that Hm" 4 HL', and the correction
on the energies of the metal orbitals is of the bonding type. In both cases, however,
e. can be defined as in equation (10-32), the only difference being in the sign (eq
is positive for antibonding, and negative for it-back bonding into empty ligand
orbitals). The energies of the metal d orbitals in a monocoordinated M-L
complex are as follows:
E(d,2 ) = e,
E(dj,) = E(d,,) = e, (10-34)
E(d,,) = E(d,2 _,2) = e.
A transition metal ion complex differs from this simple M-L example, for
we are concerned with overlap of d orbitals and we have many ligands. For an
octahedral complex, the coordinate system, shown in Fig. 10-28(A), fixes the
location of the real d orbitals. We can now employ a local coordinate system on
each ligand L, such that the metal-ligand bond is called the z'-axis. The x'-axis
is in the plane formed by z and z'. This local coordinate system is shown in Fig.
10-28(B) for ligand L2 . The polar coordinates of the ligand can be used to express
the relation between the coordinates in the primed coordinate system and those
in the unprimed coordinate system. Our concern is how to describe a d orbital,
whose position in the unprimed system is known, with the variables of the primed
Y
L3 1
L6 L
(8) (C)
FIGURE 10-28 (A) A coordinate system for a six-coordinate complex; (B) a local
coordinate system; (C) definition of 0 and 9 in Table 10-8.
coordinate system. These relationships are worked out and the results summarized
in Table 10-8. The results can be used for a complex of any geometry.
We illustrate the use of Table 10-8 by determining the relation between the
d orbitals in the primed and unprimed set for L2 . The ligands in an octahedral
complex, labeled as in Fig. 10-28, have angular polar coordinates that can be
expressed as follows:
Ligand 1 2 3 4 5 6
0 0 90 90 90 90 180
(p 0 0 90 180 270 0
Considering L 2 , we have 0 = 90 and (p= 0. Substituting these values into row

1 of Table 10-8, we find what combination of d orbitals we must take in the
TABLE 10-8. Relationships of the dOrbitals in the Primed and Unprimed Coordinate Systems of Fig. 10-25
Z2 X'Z' X'y' X'2 _ y, 2
Z2 1 (1 + 3 cos 20) 0 33sin 20 0 -( 1 - cos 20)

4 2 4
YZ - sin p sin 20 cos (Pcos 0 sin p cos 20 -cos p sin0 2 sin cp sin 20
3 1
XZ 2 cos cp sin 20 -sin 9 cos 0 -cos o cos 20 sin p sin0 2 cos 9 sin 20
2 21
XY / sin 29 (1 - cos 20) cos 2p sin 0 sin 29 sin 20 cos 29 cos 0 sin 29 (3 + cos 20)
4 2 4
X2Y 2 cos 29p(1 -cos 20) -sin 2(psin 0 1cos 2(psin 20 -sin 29 cos 0 1 cos 2(p(3 +cos 20)
4 2 4
primed system (listed across the top) to be equivalent to dz2. The result is obtained
by substituting these values for 0 and cp into the first row:
13
d,2- (1 + 3 cos 20)(z' 2 ) + 0(y'z') - /3 sin 20(x'z')
4 2
+ 0(x'y') + (1 - cos 20)(x'2 - y' 2)

4
1 2 3 2
= z' + (x' _ y, 2 ) (10-35)
2 2
2 2
The interaction energies of the orbitals z' 2 and x' _ y, with the ligand will
be given by e, and es, where the subscripts indicate that the ligand-drZ interaction
is a and that with d,2 is 6. The coefficients - /2 and + 3/2 can be considered
_ 2
as the overlap of the dZ2 orbital with d,2 and d, Since the energy is
2 2_, .
proportional to the overlap squared [equation (10-32)], we can write the energy
of the ligand-dZ2 interaction as
1 3
E(d,2) = e, + 3 e, (10-36)
We now must evaluate the effect of L 2 on the energies of all the d orbitals. We
accomplish this by substituting 0 = 90 and cp = 0 into all the expressions in
Table 10-8. We can express the result in the form of a matrix as
13
0 0 0 -
2 2
0 0 0 -1 0
0 0 -1 0 0
0 1 0 0 0
3 1
0 0 0 -
2 2
Since the columns are in the primed coordinate system, they correspond to U, 7T,
7r, 6, 6 interactions. These coefficients in the primed set produce the energies by
squaring, leading to:
1 3
E(d,2) = Ie, +3 e
E(d,,) = e,
E(d) = e. (10-37)
E(d,,) = e.
3 1
E(d,2 _,2) = e, + Ie
Substituting values of zero for 0 and (p of ligand L, into Table 10-8 yields the
matrix
0 0
0 0
1 0
0 1
0 0
Squaring the coefficients, we obtain an energy of interaction of 1 for ligand L,

with each orbital, i.e., the result given earlier by equation (10-34).
Repeating this procedure for the other ligands, we obtain:
Ligand 1 Ligand 2 Ligand 3

E(dz2) 1 3 1 3
e,+ 4e e,+ -e
4~ 4~
E(d,,)
E(dj,)
E(d,,)
E(d,2 _2) 3 1 3 1
e, + 4 e -e. + -e,,
4 46
Ligand 4 Ligand 5 Ligand 6

E(d2 2) 1 3 1 4
e, +-e,
4U 4~ 4 e, + 3e
E(dz) e,
E(dj,) e.
E(d ,) e,
E(d,, 2) 3 1 3 1
-e,+ -e,, -e.+ -e
4~ 4
You can check your result by summing each column and thereby noting that
each ligand contributes one -, two 7n,and two o types of interactions. Summing
up the contributions of the individual ligands, we obtain finally the expressions
of the energies of the d orbitals for octahedral compounds as:
E(dz2) = 3e, + 3e,
E(d,) = 4e, + 2e,

E(d) = 4e, + 2e, (10-38)
E(d,,) = 4e, + 2e,
E(d,2 _,2)= 3e, + 3 e,
Of course, the three t2g orbitals have the same energy and so do the two eg
orbitals. If the e, contribution is neglected, the difference between the eq and t2g
orbitals is 3e, - 4e., which corresponds to the A of the crystal field theory. In a
T complex, the e orbitals have energies of 8/3e, + 4 /3e3, whereas the t 2 orbitals
have energies of 4/ 3e, + 8/e, + 16/e. Note that with these parameters, AT, =
4
/ A,,,. In complexes with lower symmetry, the values of all of the ligands are
added and the d-orbital energies are calculated. Numerical values for the e,, e.,
and e, parameters are obtained from octahedral complexes. The e, values
employed in the different complexes are scaled according to the overlap integral.
The value of the model is that one set of parameters results for a particular ligand
and metal, which can be scaled for geometry and overlap to account for the
spectra of many transition metal ion complexes. The relationships of Dq, Ds, Dt,
6o, and 6n with e, and e. have been presented.(47
In closing this section, it should be emphasized that for these studies, and
other applications dependent upon band assignments, the assignments should be
verified by polarized single crystal results (see Chapter 5).
Miscellaneous 10-14 ELECTRONIC SPECTRA OF OXO-BRIDGED

Topics DINUCLEAR IRON CENTERS
Involving The electronic spectra of high spin octahedral and tetrahedral iron(III) compounds
Electronic are as expected from the Tanabe and Sugano diagrams. Three transitions are
found: 6 A1 -+ 4 T2 , 6 A i, - 4 T, and 6A 19 -- 4 A 1, [four transitions are found when
Transitions 4
E(D) is low enough in energy]; and since Dq is larger for octahedral complexes
than for tetrahedral ones, the 4 T1 and 4 T2(G) transitions occur at higher energy
in the former complexes. All of the d-d transitions are multiplicity forbidden and
are weak. However, when the electronic spectrum of the six-coordinate oxo-
bridged dimer (HEDTA Fe) 2 0 (where HEDTA is hydroxyethylethylenediaminet-
riacetate) is examined,(4 8'4 9 ) the surprising result shown in Fig. 10-29 is obtained.
The bands labeled a through d are in the correct place to be assigned to the
four d -+ d transitions, 6A1 -* 4 T1 , 6A 1 - 4 T2(G), 6A 1 -- A 14 E1, and 6A, __ 4 E(D),
of an octahedral complex. These bands are two orders of magnitude more intense
than those of a typical iron(III) complex. This intensity enhancement of spin
forbidden bands is common to spin-coupled systems because the coupling partially
relaxes the spin selection rule.
The four intense (e-h) u.v. bands are too intense for a d-d transition on a
single metal ion. The bands have been attributed( 4 8) to simultaneous d-d
transitions on the two iron(III) centers. They are coupled so that the pair excitation
is spin allowed.(4"9 The band labeled e at 29.2 x 10' cm-1 is assigned to the
simultaneous transition of a on one center and b on the other (v. + Vb =
29.4 x 10' cm-'); band f at 32.5 x 103 cm-' to a + c (32.2 x 103 cm-1); band
g at 36.8 x 103 cm- to b + b; and band h at 42.6 x 10' cm-' to b and d. There
is considerable interest in this spectral feature because of the presence of the
Fe-O-Fe group in a wide variety of iron proteins including hemerythrin and
methane monooxygenase. Resonance Raman 5 0 ) and other studies 5 1 ) of a series
of hemerythrin derivatives and model compounds has led to the conclusion that
a +c
12 - e c
C:
t
a +b
0
b
<4
a
50 40 30 20 10
10-3 V,cm-1
FIGURE 10-29 Absorption spectrum of 0.20 F aqueous solution of [(HEDTA
Fe) 20] -6H 20 at 296 K. Calculated positions of SPE excitations are indicated by arrows.
the intense near-u.v. features common to all oxo-bridged dinuclear iron(III)

compounds should be designated as oxo to iron(III) charge transfer transitions.
Similar effects have been observed on other systems( 5 ',") and have been
thoroughly studied15 3) for (NH 3),CrOCr(NH3 ),4 +. It is concluded that the
intensity enhancement arises from a vibronic, exchange-induced, electric dipole
mechanism.
10-15 INTERVALENCE ELECTRON TRANSFER BANDS
Metallomers, molecules with two or more metals, can be prepared in which the
bridging group permits the metals to exist in two different oxidation states, e.g.,
[L 5M1"-X-M"L5]+"
These "mixed valence" compounds have long been of interest because of the
intense colors they possess. For example, KFe[Fe(CN)j, Prussian blue, has a
deep color that is absent in K 3 Fe(CN), and K 4 Fe(CN),. This color has been
attributed to an intervalence transfer transition. In order to understand this
phenomenon, let us consider as an example the ruthenium pyrazine dimer(5 4):
(NH 3) 5Ru(II)-N ,QN-Ru(III)(NH3 ) 5+ 1*
(NH 3) 5Ru(II)-N QN-Ru(II)(NH3 )5±
In order to clarify several points to be made about this system, we shall

simplify our discussion by ignoring any ligands in the coordination sphere of
Ru(II) and Ru(III) except the bridging pyrazine. Furthermore, we shall label the
two ruthenium atoms as Rua"'-N N-Ru, . We shall draw a potential
energy curve( 5 5) for the molecule as a whole as we vary the Rua"-N distance,
X, in the molecule in such a way that the Rua-N plus Rub-N distance is
constant. Curve a of Fig. 10-30 results when a harmonic potential function is
assumed, i.e., V = ('/2 )kX 2 .
Next we shall construct a similar curve for the Rua"-N distance for the state
Rua"-N Q N-RuI". The equilibrium Ru,"-N distance is longer than the
equilibrium Rub"-N distance. The curve for the Rua"-N state is indicated by
curve b. We can view the two curves as defining our total system in which we
now have provision for the electron jumping from one center to the other. Eth
is the activation energy for the jump, and it corresponds in this system to the
molecule arriving at a place where the Rua-N and Rub-N bond lengths are
identical. The thermal energy required for the electron to surpass the barrier
between metals, Eth, is given(5) by
Etb = k X kX) = ET (10-39)
By the Franck-Condon principle, electronic absorption occurs without

change in nuclear coordinates, so the energy of the intervalence transfer band in
the absorption spectrum (i.e., a charge transfer transition from one Ru atom to
the other) is represented by the vertical line, Err-
The actual situation is more complex, since the coordinate X is a complicated
coordinate involving all the ligands and electrons in the complex. Depending
upon the proximity of the two metal centers and the overlap of various orbitals,
the two orbitals (one on each ruthenium) that can contain the odd electron may
mix to form a bonding and antibonding combination. This would produce the
situations depicted in Figs. 10-30(B) or (C). The quantity labeled 2H,,s in Fig.
10-30(B) is two times the resonance integral, i.e., the off-diagonal element between
the two d-orbitals which can hold the odd electron (one on each ruthenium) in
the secular determinant.( 56 ) In Fig. 10-30(B), we have shown Eth < ('/4 )E11 . In
Fig. 10-30(C), the odd electron is completely delocalized on all time scales, and
the electron absorption cannot properly be termed an intervalence transfer.
Robin and Day(" have pointed out that in Fig. 10-30(B), one of the energy
minima corresponds to an electronic wave function in which the odd electron is
mainly on one metal but, to a small extent, is delocalized onto the other metal.
They have proposed that mixed valence compounds be classified as Class I, II,
or III, depending on whether none, some, or half of the unpaired electron density
is delocalized from one metal center onto the other at any one instant.
x -x ---- 1 FIGURE 10-30 Potential

(A) energy curves for various
classes of mixed valence
compounds. The systems
vary from completely
localized systems (A) to
intermediate (B) to
V ETycompletely delocalized
2H (C). X is a coordinate
expressing the Rua'"-N
__ Eh distance as the sum of
Ru.-N and Rub-N is
held constant.
x-*
(B)
(C)
The ruthenium pyrazine dimer5 4 ) shows an intense band in the near infrared
at 1570 mp, assigned to the electronic transition shown in Fig. 10-30(A) and
labeled E,,. From the energy of this transition, a thermal rate of electron exchange
of 3 x 10' sec-' has been calculated employing a crude model and equation
(10-39). The electronic transition is described as a [2, 3] -+ [3, 2] transition.
The existence of high intensity bands in the d-d region in other complexes
has been used as evidence in support of mixed valence in molecules containing
more than one metal ion. The assignment of the band to a mixed valence species
can be supported by following the intensity of the transition as a function of the
extent of oxidation or reduction of the [n, n] oxidation state complex to form
[n, m]. A maximum is observed when [n, m] is formed, and it disappears on
formation of [m, m].
10-16 PHOTOREACTIONS
The principles discussed is this chapter and Chapter 5 are of importance to

chemists for applications other than structure determination. Fluorescence,
phosphorescence, and photochemistry all have to do with electronic transitions.
In photochemical reactions, the reactant is a molecule in a reactive excited state.
Understanding of the photochemical reaction requires an understanding of the
structure and reactivity of the excited state. In some cases, molecules that are
singlets in the ground state become reactive radicals by being excited to a triplet
state containing unpaired electrons. Often, two molecules that do not form a
complex in the ground state form a complex (called an exciplex) when one of the
molecules is in an excited state. The principles of electronic transitions studied
here and in Chapter 5 are thus important in many areas.
REFERENCES 1. C. J. Ballhausen, "Introduction to Ligand Field Theory," McGraw-Hill, New York,

1962.
CITED 2. B. N. Figgis, "Introduction to Ligand Fields," Interscience, New York, 1966.
3. M. Gerloch and R. C. Slade, "Ligand Field Parameters," Cambridge University Press,
New York, 1973.
4. a. C. K. Jorgensen, "Absorption Spectra and Chemical Bonding in Complexes,"
Pergamon Press, New York, 1962.
b. C. K. Jorgensen, "Modern Aspects of Ligand Field Theory," North Holland
5. H. L. Schlsfer and G. Gliemann, "Basic Principles of Ligand Field Theory,"
6. C. E. Moore, "Atomic Energy Levels Circular 467," Vol. 2, National Bureau of
Standards, 1952. Volumes 1 and 3 contain similar data for other gaseous ions.
7. E. U. Condon and G. H. Shortley, "Theory of Atomic Spectra," Cambridge University
8. J. S. Griffith, "The Theory of Transition Metal Ions," Cambridge University Press,
New York, 1961.
9. D. S. McClure, "Electronic Spectra of Molecules and Ions or Crystals," Academic
10. A. B. P. Lever, "Inorganic Electronic Spectroscopy," 2nd Ed. Elsevier, New York,
1984.
11. "Spectroscopic Properties of Inorganic and Organometallic Compounds," Vols. I-IV,
The Chemical Society, London, 1968-1971.
12. T. M. Dunn, "Modern Coordination Chemistry," eds. J. Lewis and R. G. Wilkins,
13. Y. Tanabe and S. Sugano, J. Phys. Soc. Japan, 9, 753, 766 (1954).
14. R. S. Drago, D. W, Meek, M. D. Joesten, and L. LaRoche, Inorg. Chem., 2, 124 (1963).
15. F. A. Cotton and M. Goodgame, J. Amer. Chem. Soc., 83, 1777 (1961).
16. D. W. Barnum, J. Inorg. Nucl. Chem., 21, 221 (1961).
17. T. S. Piper and R. L. Carlin, J. Chem. Phys., 36, 3330 (1962).
18. A. D. Liehr and C. J. Ballhausen, Phys. Rev., 106, 1161 (1957).
19. R. S. Drago, D. W. Meek, R. Longhi, and M. D. Joesten, Inorg. Chem., 2, 1056 (1963).
20. R. W. Kiser and T. W. Lapp, Inorg. Chem., 1, 401 (1962).
21. F. Basolo, C. J. Ballhausen, and J. Bjerrum, Acta Chem. Scand., 9, 810 (1955).
22. D. S. McClure, "Advances in the Chemistry of Coordination Compounds," ed. S.
Kirshner, p. 498, Macmillan, New York, 1961 and references therein.
Exercises 463
23. F. Basolo, J. Amer. Chem. Soc., 72, 4393 (1950); Y. J. Shimura, J. Amer. Chem. Soc.,
73, 5079 (1951); K. Nakamoto, J. Fujita, M. Kobayashi, and R. Tsuchida, J. Chem.
Phys., 27, 439 (1957); and references therein.
24. R. C. Fay and T. S. Piper, J. Amer. Chem. Soc., 84, 2303 (1962).
25. C. Furelani and G. Sartori, J. Inorg. Nucl. Chem., 8, 126 (1958).
26. W. Manch and W. Fernelius, J. Chem. Educ., 38, 192 (1961).
27. G. Maki, J. Chem. Phys., 29, 162 (1958); ibid., 29, 1129 (1958).
28. F. A. Cotton and E. Bannister, J. Chem. Soc., 1960, 1873.
29. N. S. Gill and R. S. Nyholm, J. Chem. Soc., 1959, 3997.
31. D. K. Straub, R. S. Drago and J. T. Donoghue, Inorg. Chem., 1, 848 (1962).
32. C. K. Jorgensen, "Energy Levels of Complexes and Gaseous Ions," Gjellerups,
Copenhagen, 1957.
33. L. Sacconi, Pure Appl. Chem., 17, 95 (1968); Transition Metal Chem., 4, 199 (1968);
R. Morassi, I. Bertini, and L. Sacconi, Coord. Chem. Rev., 11, 343 (1973).
34. C. Furlani, Coord. Chem. Rev., 3, 141 (1968).
35. a. M. Ciampolini, Struct. Bonding, 6, 52 (1969).
b. B. J. Hathaway and D. E. Billing, Coord. Chem. Rev., 5, 143 (1970).
36. B. N. Figgis and J. Lewis, "Modern Coordination Chemistry," eds. J. Lewis and
R. G. Wilkins, Interscience, New York, 1960.
37. M. Linhard, H. Siebert, and M. Weigel, Z. Anorg. Allgem. Chem., 278, 287 (1955).
38. C. J. Ballhausen and H. B. Gray, Inorg. Chem., 1, 111 (1962).
39. H. B. Gray and C. R. Hare, Inorg. Chem., 1, 363 (1962); C. R. Hare, I. Bernal, and
H. B. Gray, Inorg. Chem., 1, 831 (1962).
40. A. S. Brill, B. R. Martin, and R. J. P. Williams, "Electronic Aspects of Biochemistry,"
ed. B. Pullman, Academic Press, New York, 1964.
41. D. R. McMillin, R. A. Holwerda, and H. B. Gray, Proc. Nat. Acad. Sci., 71, 1339 (1974).
42. W. S. Caughey et al., J. Mol. Spect., 16, 451 (1965) and references therein.
43. D. A. Rowley and R. S. Drago, Inorg. Chem., 7, 795 (1968); ibid., 6. 1092 (1967).
44. a. A. B. P. Lever, Coord. Chem. Rev., 3, 119 (1968).
b. R. L. Chiang and R. S. Drago, Inorg. Chem., 10, 453 (1971).
45. C. D. Burbridge, D. M. L. Goodgame, and M. Goodgame. J. Chem. Soc. (A), 1967, 349.
46. a. C. E. Schaffer and C. K. Jorgensen, Mol. Phys., 9, 401 (1965).
b. E. Larsen and G. N. LaMar, J. Chem. Educ., 51, 633 (1974).
47. 1. Bertini, D. Gatteschi, and A. Scozzafara, Inorg. Chem., 15, 203 (1976).
48. H. J. Schugar, G. R. Rossman, J. Thibeault, and H. B. Gray, Chem. Phys. Letters, 6,
26 (1970).
49. L. Lohr, Coord. Chem. Rev., 8, 241 (1972).
50. 1. Sandere-Loehr et al. J. Am. Chem. Soc., 111, 8084 (1989).
51. R. C. Deem et al. J. Am. Chem. Soc., 111, 4688 (1989).
52. A. E. Hansen and C. J. Ballhausen, Trans. Faraday Soc., 61, 631 (1965).
53. H. U. Gbdel and L. Dubicki, Chem. Phys., 6, 272 (1974) and references therein.
54. C. Creutz and H. Taube, J. Amer. Chem. Soc., 91, 3988 (1969).
55. N. S. Hush, Prog. Inorg. Chem., 8, 391 (1967) and references therein.
56. B. Mayoh and P. Day, J. Chem. Soc., Dalton, 1974, 846.
57. M. B. Robin and P. Day, Adv. Inorg. Chem. Radiochem., 10, 248 (1967).
1. Refer to the Tanabe and Sugano diagrams in Appendix D. For octahedral Cr(III) EXERCISES
and a ligand with Dq/B of 1, how many bands should occur in the spectrum? Label
these transitions and list them in order of increasing wavelength.
2. In complexes with weak field ligands (Dq/B = 0.7), octahedral Co 2 exhibits a spectrum
with three well-separated bands. Make a tentative assignment using the Tanabe and
Sugano diagrams and list the assignments in order of decreasing frequency. Would
the spectrum of a strong field complex be any different? Describe the spectrum you
would expect for a strong field complex.
3. A nickel complex NiR 4 C 2 has an absorption spectrum with peaks that have e values
of around 150. R and Cl occupy similar positions in the spectrochemical series. Are
the chlorines coordinated?
4. Two different isomers of Co(NH 3 )4 (SCN) 2 * were separated. How could you determine
whether the SCN groups in both were bonded through the sulfur? If both isomers
were coordinated through sulfur, how would you determine which is cis and which
is trans? (Hint: -SCN is near Cl in the spectrochemical series, while -NCS- creates
a stronger field; Co(NH 3 )4 Cl 2 + is easily prepared.)
5. Using the Tanabe and Sugano diagrams, assign the following spectra of six-coordinate
aquo species [except for (A)].
K CoF
3V 6
10,000 15,000 20,000 25,000

Cm-
(A)
2 4
Cr (d )
5 - 0.10 M in 0.75 M H2 SO4
E
5,000 10,000 20,000 30,000 35,000

Exercises 465
2.0 -
0Fe 6)(CO
C~ /,' 0.077 M Fe(C 104) 2
01
5,000 10,000 20,000 30,000 35,000
2
_ Mn +(d5)
- 0.36 M Mn(CI0 )
4 2
.e . . . . . I..
. I I.
5,000 10,000 20,000 30,000 35,000
0.143 M TiC
3
0. p p % i i I I A
I I a iI A
5,000 10,000 20,000 30,000 35,000
cm-
cm-1
2
Ni + (dl)
0.0025 M Ni(C10 )
C 42
5,000 10,000 20,000 30,000 35,000
CU2+ (d9)
0.0030 M Cu(CI0 )
4 2
0 i 1 ' III ' ' 11
5,000 10,000 20,000 30,000 35,000

cm-
6. Using the spectrum in problem 5(H), calculate the value of 1ODq and B for H 2 0.
7. Given the following information for d 3 [see, e.g., A. B. P. Lever, J. Chem. Ed., 45, 711
(1968)]: 4 A 2g - 4 T2g = 1ODq and 'A 29 - 4 Ti,(F) = 7.5B' + 15Dq - 1/2 (225B' 2 +
10ODq 2 - 180B'Dq) 2; calculate the value of 1ODq and #l(B'/B where B = 1030 cm -)
for:
a. H 2 0, using the spectrum in problem 5(B).
2
b. C 2 0 4 , using the following spectrum:
0.0050 M K 3Cr(ox)
3
5,000 10,000 15,000 20,000 25,000 30,000 35,000

cm-1
c. The value of 1ODq for H 2 0 toward Ni 2 + is 8500 cm-', and # = 0.88. Compare
your results in part (a) with these values and offer an explanation. Also compare
your results toward Cr(III) for water with those for oxalate.
2
d. Given that "A2. -* 'Tg(P) is given by 7.5B' + 15Dq + 2 (225B'2 + 1OODq
180B'Dq) 2, calculate the frequency of this band in Cr(C 2 04)3 3 -
8. The ion [Ni(pyridine)4 (H 2O) 2 ] 2 + has d-d absorption bands at 27,000, 16,500, and
10,150 cm-'. No low symmetry splitting is observed. Treating it as an octahedral
complex, determine 1ODq. Compare this value with the average (rule of average
environment) of the values predicted from the two six-coordinate complexes. The Dq
values for the six-coordinate complexes can be predicted from Table 10-6.
9. Why are octahedral Mn2+ complexes (weak field) much less intensely colored than
those of Cr 3 +
10. The electronic spectrum of trisoxalatochromium(III) doped into a host lattice of

NaMgAl(C 20 4 )3 -9H 20 has been reported. The ground state is 4 A 2g if octahedral
symmetry is assumed. The lowest excited states (not m.o.'s) for octahedral symmetries
are then 2Eg, 2 Tjq, 2 T2g, and 4 T2 ,. The observed bands and extinction coefficients are:
17,500 cm - ' r 40
23,700 cm -1 e= 67
14,500 cm - ' r = 2.6

15,300 cm -1 e = 2.0
20,700 cm - ' e = 0.3
Exercises 467
a. The above transitions have been shown to be electronically allowed. Why is this
spectrum inconsistent with octahedral symmetry? (Multiplicity-forbidden transi-
tions have low intensities; generally, - is less than 5.)
b. Actually, the spectrum is consistent with the true symmetry of the molecule, D3-
Lowering the symmetry causes the following:
Oh D3
A 2g A2
TI A 2 +E
T, A 1 +E
E, E
If the spectrum is taken with light polarized perpendicular and parallel to the trigonal
axis, all the absorptions except the one at 17,500 cm -1 occur with perpendicularly
polarized light, but only the bands at 17,500 and 15,300 cm -' are present with parallel
polarized light. Given that splittings of the doublet octahedral states in D3 symmetry
are unresolved and that 2 Eg is lower in energy than 2 Tig, assign the transitions using
the D3 excited states. You must explain your choices. (Remember that some doublet
bands represent unresolved multiplets and will consequently correspond to more than
one transition.)
11. In Re 2C] 8 2 ~ (see problem 6, Chapter 3) the transition of an electron from the b2,
orbital to the bl orbital is a d-d transition in a molecule with a center of inversion.
Is it allowed? Explain.
0
12. The following is the splitting of the state energies of a high spin 3d2 ion in an , field:
3
A 2q
(*)
3
Tig
a. Calculate the free ion ground term symbol (*). (Note the T and A states written
above will have the same spin multiplicity as the free ion.)
b. What would this diagram look like if we considered spin-orbital (;,L -S) effects? (Do
only the T,, state.) Show the splitting and label each level with its J value and its
degeneracy.
c. Show the splitting pattern for a d2 case for all J states above and label with
appropriate Mj values. Show the effect of a magnetic field on the levels.
13. a. Of the states arising from an ion with a configuration of two 2 p electrons, only
one is a spin triplet. What is the term symbol of this triplet? Show all work.
b. Now consider spin-orbit coupling. What J states will this triplet give rise to? Which
will have the lowest energy?
14. a. Determine the irreducible representations for the terms arising from the splitting
of a gaseous ion 4F state by an octahedral ligand field.
b. Using the method of descent in symmetry, determine the spin multiplicity of the
states in part a.
15. Determine the irreducible representations of the states into which the 2
T,, level of
Fe(CN), 3 - is split by spin-orbit coupling.
16. a. List the operations of the double group D3 ' that are obtained by combining the
operations of D 3 with the operation R. Make sure your list has the properties of
a group.
b. How many classes are in D 3 '? What are the operations in each class? You may wish
to utilize the fact that equivalent symmetry operations are members of the same class.
17. Indicate the S quantum numbers and calculate the energies of the states arising from
the magnetic coupling of S1 = 2 and S2 = 5/2 metal centers.
Magnetism 11
11-1 INTRODUCTION
In this chapter we will consider certain aspects of magnetism that are critical to
an understanding of the nmr and esr spectra of transition metal ion complexes.
Magnetic effects arise mainly from the electrons in a molecule because the
magnetic moment of an electron is about 10' times that of the proton. In the
chapters on nmr, we have discussed the electron circulations of paired electrons
that give rise to diamagnetic effects. When there are unpaired electrons in the
system, we observe magnetic behavior that is related to the number and orbital
arrangement of the unpaired electrons. The magnetic behavior is determined by
measuring (vide infra) the magnetic polarization of a substance by a magnetic
field. Various types of behavior are illustrated in Fig. 11-1. It is convenient to
N-S N S N=S
(A) (B) (C)
FIGURE 11-1 (A) Magnetic field lines of flux (i.e., contour lines of constant field
values) in vacuum; (B) the lines of flux for a paramagnetic substance in a field; (C)
the lines of flux for a diamagnetic substance in a field.
define a quantity called magnetic induction, B, in order to describe the behavior

of substances in a field.
B = Ho + 47r N (11-1)
Here Ho is the applied field strength and M is the magnetization, i.e., the intensity
of magnetization per unit volume. When we divide* by H,, equation (11 -2) results:
* We assume in an isotropic system that the directions of Hi0 and M are coincident. Thus,
though we cannot divide a vector by a vector, we can factor out the directional property and perform
the division, leaving an equation in which only the magnitudes appear. 469
470 Chapter 11 Magnetism
B M'
= I + 47 = 1 + 4x (11-2)
Ho Ho
Here M'/H, is given the symbol X,and is referred to as the magnetic susceptibility
per unit volume. The volume susceptibility is thus related to the magnetization by
M'
-
Ho (dimensionless)
B/H, is the permeability of the medium and is the magnetic counterpart of the
dielectric constant. Dividing x, by the density of the substance, d, produces the
gram susceptibility, Zg:
X. (cm3/gram) (11-3)
Multiplying Zg by the molecular weight produces a molar susceptibility, Z:
Z, x MW = z (cm"/mole) (11-4)
The value of Z is negative for a diamagnetic substance and positive for a

paramagnetic one. In an ordered crystal, the susceptibility may be anisotropic,
i.e., represented by a tensor with several components. We shall discuss five types
of magnetic behavior: diamagnetism, paramagnetism, superparamagnetism, fer-
romagnetism, and antiferromagnetism. The behaviors corresponding to various
classifications are described in Table 11-1 and superparamagnetism will be
discussed in a later section. The latter two types of behavior can be checked by
studying the field dependence of Z. The behavior of the susceptibility as a function
of temperature is also quite characteristic for these different substances. This is
illustrated in Fig. 11-2.
TABLE 11-1. Various Types of Magnetic Behavior

Field Dependence
Type Sign Magnitude of X Origin
Diamagnetism 10-6 emu units Independent Field induced, paired electron

circulations
Paramagnetism + 0 to 10-4 Independent Angular momentum ofthe electron
emu units
Ferromagnetism + 10-4 to 10-2 Dependent Spin alignment from dipole-dipole
emu units interaction of moments on adjacent
atoms, 11
Antiferromagnetism + 0 to 10 -4 Dependent Spin pairing, 1 ,from dipole-dipole
emu units interactions
FIGURE 11-2 Temperature

dependence of
ferromagnetic, paramagnetic,
and antiferromagnetic
behavior.
Ferromagnetic
I Paramagnetic
I Antiferromagnetic
TN Tc
The temperature at which the maximum occurs in the plot of antiferromag-

netic behavior is referred to as the N el temperature. The temperature at which
the break occurs in the ferromagnetic plot is called the Curie temperature. Many
compounds which, in the solid state, exhibit paramagnetic behavior around room
temperature exhibit slight ferromagnetic or antiferromagnetic behavior below
liquid helium (4.2 K) temperature.
11-2 TYPES OF MAGNETIC BEHAVIOR
Diamagnetism
As mentioned earlier, diamagnetism arises from field-induced electron circulations
of paired electrons, which generate a field opposed to the applied field. Thus, all
molecules have contributions from diamagnetic effects. The diamagnetic suscep-
tibility of an atom is proportional to the number of electrons, n, and the sum of
the squared values of the average orbital radius of the ith electron r,:
- Ne2 n~ (1
A .2 -2.83 x 1010 2 (11-5)
Larger atoms with more electrons have greater diamagnetic susceptibilities than
smaller atoms with fewer electrons. The molar diamagnetic susceptibility of a
molecule or complex ion, x,can be obtained to a good approximation by summing
the diamagnetic contributions from all of the atoms, yA, and from all of the bonds
in functional groups, ZB.
z =
YZ' _ Z' i (11-6)
J
TABLE 11-2. Pascal's Constants

Atoms, ZA Bonds, XB
ZA(X 10-6 XA(X 10 6

3 3
Atom cm 3 mole-') Atom cm mole --1) Bond XB( X 10-6cm mole~')
H -2.93 F -63 C=C + 5.5

C -6.00 Cl -20.1 C--C +0.8
C (aromatic) -6.24 Br -30.6 C=N +8.2
N -5.57 1 -44.6 C=-N +0.8
N (aromatic) -4.61 Mg2+ ~5 N=N +1.8
N (monamide) -1.54 Zn2+ -15 N=O +1.7
N (diamide, imide) -2.11 Pb2+ -32.0 C=O +6.3
o -4.61 Ca2+ -10.4
02 (carboxylate) -7.95 Fe2+ -12.8
S -15.0 Cu 2 + -12.8
P -26.3 Co 2 + -12.8
Ni 2 + -12.8
The values of ZA and Z, are referred to as Pascal's constants, and some common
ones are listed in Table 11-2. The calculation of x is illustrated here for pyridine
and acetone.
CSHSN
3
SUM OF CONTRIBUTIONS TO Z (x 10- 6 CM MOLE-')
5 x C (ring) = -31.2
5 x H = -14.6
1 x N (ring) = -4.6
3
Z = Y_Zi + Y_ y, =-50.4 x 10-6 cm mole- 1
i j
The functional groups are accounted for by using the ring values for carbon and
nitrogen, so Y XB, equals zero.
(CH3)2C = 0
3
SUM OF ATOM CONTRIBUTIONS (X 10 6 CM MOLE-')
3 x C = -18.0
6 x H = -17.6
1 x 0 = -4.6
6 3
SUM OF BOND CONTRIBUTIONS (x 10 CM MOLE )
1 x C=O +6.3
3
XA YB = -33.9 x 10-6 cm mole
For a transition metal complex, one can measure only the net magnetism,
XMEAS, which is the sum of the paramagnetic, XPARA' and diamagnetic, ZDIA>
contributions.
ZPARA = XMEAS - ZDIA (11 -7)
Thus, to obtain the paramagnetic susceptibility, the diamagnetic susceptibility

must be subtracted from the net susceptibility. This can be accomplished by: (1)
using the values for Pascal's constants reported in Table 11-2 to estimate XDIA;
(2) by measuring the diamagnetic susceptibility of the ligand and adding that of
the metal from Table 11-2 to obtain xDIA; or (3) by making an analogous
diamagnetic metal complex and using its value as an estimate of /DIA'
Paramagnetism in Simple System where S = '2

The paramagnetic contribution to the susceptibility arises from the spin and
orbital angular momenta of the electrons interacting with the field. First, we shall
consider a system that is spherical, contains only one electron, and has no orbital
contribution to the moment. The magnetic moment, Ii, associated with such a
system is a vector quantity given by equation (11-8):
sf -g#S (11-8)
where S is the spin angular momentum operator, g is the electron g-factor

discussed in Chapter 9, and #lis the Bohr magneton of the electron, also discussed
in Chapter 9 (#l = 0.93 x 10-20 erg gauss-).
The Hamiltonian describing the interaction of this moment with the applied
field, H, is given by
H = - iH = g#S H (11-9)
This Hamiltonian, operating on the spin wave functions, has two eigenvalues
(see Fig. 9-1) with energies given by
E mg#H with m, = i (11-10)
with an energy difference

AE = g#H (11-11)
When H is about 25 kilogauss, AE for a free electron with g = 2.0023 is about

2.3 cm-', which is small enough compared to kT(205 cm 1at room temperature)
that both states are populated at room temperature, with a slight excess in the
ground state.
The magnitude of the projection along the field direction of the magnetic
moment, p, of an electron in a quantum state n is given by the partial derivative
of the energy of that state, E., with respect to the field, H, as shown in equation
(11-12):
0E
y, = - H m,g#l (11-12)
In order to determine the bulk magnetic moment of a sample of any material,

we must take a sum of the individual moments of the states weighted by their
Boltzmann populations.
The Boltzmann factor for calculating the probability, P., for populating
discrete states having energy levels E. at thermal equilibrium is given by equation
(11-13):
N exp
N ---- x" (11 13)
exp kT
Here N, refers to the population of the state n, while N refers to the total
population of all existing states. We have a wave function for each state, and we
use the term 'level' to indicate all of the states that have the same energy. The
population-weighted sum of magnetic moments over the individual states, which is
the macroscopic magnetic moment, M, then, is given for a mole of material by
equation (11- 14):
M= N I pP (11-14)
where N is Avogadro's number. Substituting equation (11- 13) for P in equation

(11-14) produces (11-15) for an S = '/2 system.
+1/2 (- En
N_ Ynp exp"
m - -/2 kT
M = 12 (11-15)
m, - Z
1/2 exp(n(k T
Substituting equation (11-12) for p, and equation (11-10) for E., and summing
over m, = i '/2, produces:
g#H -gplH
Ng 2kT
M -O
2 epg#H +ep-
( 2kT)11-16)
2kT)
gfH
(
e 2kT + exp 2kT
When (gpH/k T) < 1 (gpH equals - cm 1 for g = 2.0 and common fields of
5000 to 10,000 gauss, and kT is 205 cm-1 at room temperature), we can introduce
the following approximation:
exp 2kT ( + 2kT

( 17)
Substituting equation (11-17) into (11-16) and simplifying leads to
M Ng=~kH (11 18)

4kT
Since the molar susceptibility is related to the moment by
M M (11-19)
H
we can write
(11-20)
4kT
Equation (11-20) is the so-called Curie law, and it predicts a straight-line

relation between the susceptibility and the reciprocal of temperature, giving a
zero intercept; i.e., Z - 0 as T approaches infinity. This type of behavior is usually
not observed experimentally. Straight-line plots are obtained for many systems,
but the intercept is non-zero:
C
T- 0 (11-21)
Equation (11-21), where C = Ng 2 # /4k and 0 corrects the temperature for the
2
non-zero intercept, describes the so-called Curie-Weiss behavior. It is common

to have a non-zero intercept in systems that are not magnetically dilute (i.e., pure
solid paramagnetic material). In these systems, interionic or intermolecular
interactionso'1 cause neighboring magnetic moments to become aligned and
contribute to the value of the intercept.
For the case of molecules with no orbital angular momentum, one commonly
sees equation (11-20) written as
N g2fl 2
X= - S(S+ 1) (11-22)
3kT
with units of Bohr magneton, BM.
This reduces to equation (11-20) for S = '/2 and yields the so-called spin-only
magnetic susceptibilities of complexes containing different numbers of unpaired
electrons.
Both Z and M are macroscopic properties. In describing the magnetic
properties of transition metal complexes, it is common to employ a microscopic
quantity called the effective magnetic moment p It is defined as follows:
p = (31)1/2 (ZT)1/2 = 2.828(yT)1/2 (BM) (11-23)
The susceptibility employed in this equation is ZPARA described above. Equation

2
(11-23) can be obtained by replacing g S(S + 1) in equation (11-22) by P f2 and
solving for pef. Thus, any effects that tend to make S not a good quantum number
become incorporated into the g-value (Recall the variability of the g-value in epr,
TABLE 11-3. Spin-Only Magnetic Moments for Various

Numbers of Unpaired Electrons
Number of Unpaired
Electrons S pi,(spin-only)(BM)
1 1/2 1.73
2 1 2.83
3 3/2 3.87
4 2 4.90
5 5/2 5.92
6 3 6.93
7 7/2 7.94
Chapter 9.) Equation (11-24) can be used to calculate the spin-only magnetic
moments for various values of S:
peff (spin-only) = g[S(S + 1)]1/2 (BM) (11-24)
where g equals 2.0 for an electron with no orbital angular momentum.

The spin-only results in Table 11-3 are obtained for various numbers of
unpaired electrons. In many transition metal ion complexes, values close to those
predicted by the spin-only formula are observed.('-9) However, in many other
complexes, the moments and temperature dependence of the susceptibility are at
variance with these predictions. Other effects are operative, and a more complete
analysis is in order.
11-3 VAN VLECK'S EQUATION
General Basis of the Derivation

In this analysis, we will introduce orbital contributions and also anisotropy in
the magnetic susceptibility for low symmetry molecules. Defining the principal
molecular axis as the z-axis, we can write the necessary part of the Hamiltonian
including these additional effects as:
S= 2 $ + #f(L
+ g5)H (11-25)
where L and 5 are operators with x, y, and z components. In this chapter, g,

will be used when referring to the free electron g-value. The first term on the
right of the equality sign describes the spin-orbit coupling (i is the spin-orbit
coupling constant) and is seen to be field independent. The other term sums the
spin and orbital contributions to the electron moment. [Note its resemblance to
equation (11-9).] In using this Hamiltonian, we have to worry about what basis
set to employ. In a free ion, for a d' case with L - S coupling ignored, the complex
functions + 2>, 1+ 1>, etc., are a good choice for they are already eigenfunctions
of h. Hence, when the full matrix with elements (p, ILz + gS2| p.>#H is
evaluated, there will be no off-diagonal elements.
11-3 Van Vieck's Equation 477
In a d' complex, with the ligands defining the x, y, and z axes, the orbitals
are a convenient basis set. Now we have, for example:
1
dx,= (d+12 -- d-2)
i/2
and
1
d,2 _),2 (d+2 + d-2)
S2
In this basis set, the off-diagonal elements will be non-zero; for example,
1
(dxy IlzI ,|d _,,2 = ((Ed+2| I Id+2) +(<d+121 LzId- ( d- | Id+12)
<d- 2 |14Id-2>]
= [2+ 0 +0 -(-2)]= -2i

2i
The Sz contribution to this matrix element will be zero, since
(dx,|$1 ,d_*)= <d,,|dX2_ >)+ $Z I+> = (0) 2 0
The non-zero off-diagonal elements account for a distortion of the ground state
wave function by the applied field [we worked out the above matrix elements
for Lz and $, but the full Hamiltonian is #l(L+ g,5) -H]. This distortion is
accomplished by mixing in appropriate excited states. The diagonal elements are
called the first-order Zeeman terms, and the off-diagonal elements give rise to
second-order Zeeman terms. If there were no off-diagonal terms, all of the diagonal
matrix elements would be to the first power in H and the resulting energies would
be first order in H.
Off-diagonal elements connecting states of very different energies are
generally small compared to the energy difference, so this problem is generally
treated with perturbation theory. We saw in the discussion of the Ramsey equation
(Chapter 8) that this approach gave rise to terms of the general form
msn Em - E
where Op is an abbreviation for an unspecified operator. The field-induced mixing

of excited states was used as an interpretation of the paramagnetic contribution
to the chemical shift. In the present case, we obtain from perturbation theory a
term that looks similar:
(11 26)
j#k i Ei -- E.j
E
When the ion configuration becomes other than d', the advantage of the
perturbation treatment is seen, for the full matrix would be large.
For a weak complex, the basis set for a full matrix evaluation would involve
using the wave functions that resulted after a weak field approximation in the
crystal field analysis of the electron-electron repulsions. For a strong field complex,
the real d orbitals provide a good basis set for the complex. Thus, we see that
the relative magnitudes of the factors influencing the d orbital energies are
important in determining the best basis set. We can list some common, rough
orders of magnitude as follows, using C.F. to abbreviate the crystal field:
(A) Weak field, first row transition metal ion:
e2/rij > C.F. > low-symmetry C.F. perturbation> L-S
104 cm 1 104 cm- 1 102 to 103 cm -1 102 cm -I
(B) Strong field, first row:
C.F. > e 2 /rij 3> C.F. perturbation> L S

5 x 104 cm -1 3 x 10' cm 1 103 cm -1 102 cm -1
(C) Third row:
C.F. > 2LS > e2 r
104 cm -1 103 cm - 1 103 cm -1
(D) Lanthanides:
AL -S > e 2 /ri > C.F.
5 x 103 cm -1 5 x 103 cm- 1 103 cm-1
The magnetic field effect is about 1 cm-'.
In our analysis so far, we have not taken spin-orbit coupling (the AL S
term) into account. For first row transition metal ions, this is accomplished by
adding the effects of L - S to the energies as a perturbation on their magnitude.
This is a good approximation only when AL -S is small compared to electron-
electron repulsions and crystal field effects. The diagonal L - S matrix elements
are evaluated in the real orbital basis set and added to the energies as corrections.
When spin-orbit coupling is large, this perturbation approach is not appropriate.
For example, d2 and d, ± (signs refer to the electron m, value) have the same
mi value (3/2) and are mixed by L - S.
Derivation of the Van Vieck Equation

This very general discussion of how to proceed in a crystal field evaluation of
the effects of the Hamiltonian in equation (11-25) on the molecule or ion of
interest is sufficient for our purposes. We shall now return to a discussion of the
influence of these factors on the magnetic moment. When we list the contributions
to the energy of a given state, n, from the factors discussed in the earlier section
for S = 1/2 systems, in terms of the field dependence of the effects, equation (11-27)
results:
E, = En" + 2
HE ") + H 2 E, ) (11 -27)
SL S first-order Zeeman second-order Zeeman
(diagonal terms) (off-diagonal terms)
Recalling that the projection of the magnetic moment in the field direction is
given by -- 8E,/8H [equation (11-12)], we see that the first term, E."', makes
no contribution to the moment of a given state; the second term makes a
contribution that is independent of the field strength; and the third term makes
a field-dependent contribution. The E " term in equation (11-27) is the same
term that we had in the Curie law derivation, except that the orbital momentum
is now included. The magnitude of the second-order contribution will depend
upon Ei - E . It can be very large when the electronic excited state is close in
energy to the ground state and has correct symmetry.
In order to determine the influence of these effects on the susceptibility, we
return to the earlier Curie law derivation and rewrite equation (11-15) by replacing
exp (- E,/kT) with
E - HE - H2E w +. . (. HE (') -E( 0 ))

exp kT kT kT
(11-28)
Also, we let
- BE
OH=
p "=- E - 2HE (2)(11-29)
Making these substitutions into equation (11-15) leads to:
Y( - E (' - 2HE Q) 1- H " exp " O)

M N kT kT (11-30)
exp E(O) HE ")
n kT kT)
Limiting this derivation to paramagnetic substances, this equation must yield

M = 0 at H = 0 and, in order for this to happen, the following must be true:
- E.") exp E(O) 0 (11 31)
Expanding the numerator, and neglecting terms higher than E (2 ) as well as the
E 2) En" product in equation (11-30), and recalling that X = M/H [equation
(11-19)], we obtain from equations (11-30) and (11-31):
[ (E. 1)2 -E ()
"k - 2En m exp
X iN " (11 32)
exp kT
where En" has contributions from AL - S, etc. The En") term is always zero for
the ground level; for a higher energy state, the quantity giving rise to this energy
term in the absence of a field is substituted for E.4". The term E,(') contains the
mg#H and other first-order contributions, and Ent 2) has the contributions from
the second-order Zeeman term.
Thus, the susceptibility is determined by taking a population-weighted

average of the susceptibility of the level. An r-fold degenerate level has r component
states, each of which must be included in the summations of equation (11-32). Its
use will become more clear by working out some examples.
Applicationof the Van Vleck Equation

We shall demonstrate the use of the Van Vleck equation [equation (11-32)] by
applying it to the ground state of a free metal ion with quantum number J
(Russell-Saunders coupling applies). In all of the examples worked out in this
section, it is important to appreciate that all we are doing is taking a
population-weighted average of the individual moments of the levels. The 2J + 1
degeneracy is removed by a magnetic field, and the relative energies of the resulting
levels are given by msg#H. We are considering only the ground level En" and
E , which are taken as zero. (In doing a Boltzmann population analysis, the
zero of energy is arbitrary; we set the energy of the ground level in the absence
of H[i.e., E,, 0)] at zero for convenience.) Equation (11-32) becomes
+ mi 2g 2#2
mj Y kT(2J + 1)
n-0
Ng 2 2
j2 + (J - 1)2 + .. +( J + 1)2 + (-j)2
kT ( 2J +1
Ng 2f32 J(J + 1)(2J + 1) Ng 2 # 2 J(J 1) (11-33)

kT 3(2J + 1) 3kT
Written in terms of pen" we obtain for the free ion
pe-f = g[J(J + 1)]'/2 (BM) (11-34)
For a free ion, following the Russell-Saunders coupling scheme, we give without
derivation(2 ) the expression for the g-value as:
I + S(S + 1) - L(L + 1) + J(J + 1) (11-35)

2J(J + 1)
We see from this equation that, in the free ion, contributions to pef arise from
both the spin and the orbital angular momenta. Furthermore, when L = 0, then
J = S. Then g = 2.00 and equation (11-34) reduces to the spin-only formula given
in equation (11-24).
The next example selected to illustrate the use of equation (11-32) is a
Ti" (d') complex.>"The splitting of the gaseous ion terms by the crystal field,
spin-orbit coupling, and the magnetic field is illustrated in Fig. 11-3. The
expressions for the energies given in the figure are obtained from the wave
functions resulting from a weak crystal field analysis by operating on them with
the ^L - 5 and #(^- + g,$) H operators of equation (11-25). Since 1ODq is generally
large in an octahedral complex, we can ignore the 2 E state in evaluating the
susceptibility with equation (11-32). We shall discuss this entire problem by
starting with the ground 2 T 2 level and numbering the states 1 to 4 in order of
I FIGURE 11-3 The splitting

Oq of the gaseous ion 2 D state
2__ Enrge by an 0 ,, field, by A~ S and
(10) '\ ) A + flH + 4 '82H 2 by a magnetic field. The
(2) F 3 X degeneracies of the levels
-(1) 4 p 2 H2 are indicated in parentheses,
2T2" 3 Xand the energies are listed
2T2 __3 A
X-X/2on
(6) the right.
(6) -A/2
r(2) -A/ 2H2
(4)s _2L 4 /32H2
2 3 A
0
Gaseous h Field ALS H
ion
increasing energy. The Eo terms for states I to 4 are -the2, g u2, oandnD
respectively (see Chapter 10). The Eb terms are 0, 0, and +LSH, fdH,
respectively, while the E.(2 terms are - 4/ 3 (fl2 H2 /,), 0, + 4 3 (3 2 H2 A), and
4 2
+ 3 (fHb respectively. Substituting these values into equation (11a32) and y),
multiplying each term by the degeneracy of the corresponding level produces:
~ 2[(,-4 + (2)( )- y exp (2T]+ 2[(i-T) (2)(0) exp (~)
+
+ad 2 (4~)(,2] x (_)+ [(,2) ()(,2] x()
te(2)
eri are liste
Recalling that e
(NE2 3kT)ter 2 s we obtain:
+2~
ep2T+x 2kT +exp kT
8= expe- (2)8]
32 31
kT I ( 2 kT
where # is the Bohr magneton. Note that our analysis predicts that the Curie
law will not hold. As T approaches infinity, pef 2 approaches zero, As T becomes
small, pe, 2 becomes small; and as T approaches zero, the equations no longer
apply because g#H ~ kT. As A approaches zero, peff 2 approaches 3. Finally, as
T approaches zero, we have the very interesting result that a system with one
unpaired electron has zero susceptibility. The result arises because the spin and
orbital contributions cancel. These predictions are confirmed by experiment.
3 In our analysis of this problem, we have ignored any contributions from the
2
E9 excited level. However, the above approximation is valid for many magnetic
2
applications. In the more sensitive epr technique, one can detect the contribution
(4) from the excited state to the g-value (see Chapter 12). With spin-orbit coupling,
4 D
A2
the ground level (F,) and the F, excited level from 2 Eg (F, x F, = F,) can mix,(4)
2 changing the g-value from 4#H/32 to 4;/A + 4#H/32, where A is 1ODq. The
(2)
second-order Zeeman term mixes the ground level with the excited level, and the
extent of mixing depends upon A.
FIGURE 11-4 The splitting
of the 4A 2 state by a
The next system (9) that we shall consider is chromium(III) (d3 ). An octahedral
tetragonal field D. (D is the field gives rise to a 'A 2 ground state and 4 T 2 and 4 T, excited levels, as shown
tetragonal splitting or for the quartet states in the Tanabe and Sugano diagrams. Since 4 T 2 is about
zero-field splitting 18,000 cm-' higher in energy than 4A 2, its contribution to the susceptibility can
parameter.) be ignored. Since the ground state is orbitally singlet (A), there is no orbital
contribution to the magnetic susceptibility (vide infra). The magnetism is predicted
with the spin-only formula and S = 3/2. Next, we shall consider the effect of a
tetragonal distortion on chromium(III). This removes the degeneracy of the
m, = '/2 and the m, = + 3/2 states as shown in Fig. 11-4. The splitting by the
tetragonal component is described by the parameter D. Since this splitting exists

in the absence of a field, it is one of the many effects that are referred to as a
zero-field splitting. For the case of an axial zero-field splitting, one can represent
this with the Hamiltonian, DSz 2 . The susceptibility for this system when the
applied field is parallel to the principal molecular axis is obtained by inserting
values for E,,(", which equals ('/ 2)gz# and (3/2 )gz#, respectively, for the two levels
shown in Fig. 11-4. With E,(') given a value of zero for the lower energy level
and a value of D for the higher one, we have:
2__g_# 2( gj)# -D] -DN

Sexp(k)Txp (0) + kT DkT 2 exp(0) + 2 exp (kT)
g4kT#21 1 + 9 exp -D\

4kT [I + exp( )]
Thus we see that when (D/kT) < 1, which is true for a very small distortion or
at a very high temperature, the expression for Zz reduces to (5/ 4 )Ngz2 2 /k T,
while the spin-only formula for S = '/2 results as T approaches zero or D becomes
very large.* To calculate the powder average susceptibility, x. and Z, must be
evaluated using the x and y components of L and 5 in equation (11-25) with the
D$ term added. The anisotropy in y can be calculated this way.
As a final example, we shall consider a nickel(II) complex with a small
tetragonal distortion. The splitting is shown in Fig. 11-5.
* Experimentally, D is approximately 0.1 cm-' for pseudooctahedral Cr" from a spin-orbital

mixing in of excited states.
3
T1
3
F ,-'' Energies
(2) , +PH,+ D
FIGURE 11-5 Splitting of a
3A 2 gM=-1 -#H, + D nickel(II) ion in a tetragonal
(3) D s field.
Ms= 0
(1)
Gaseous Oh Tetragonal H
ion field field
(-g,) (-D) (g #)2 -- D)

(0) exp (-0) + exp + exp
ZI kT kT kT kT
N +2exp(-D)
kT
2 2
2_Ng exp (-D)
kT - D
1+2exp
( kT
When D < kT, the following expression results:
2Ng 2fl2 1 D g2# 2

--1 kT) 2Ng,2I D
2-2D) 3kT kT
kTkT1
1+2~~
kT
For (NH 4 )2 Ni(S0 4 )2 -6H 2 0, the experimental value(4 ) of g is 2.25 and D is -2.24
cm', giving a value of Z1 of 4260 x 10-' cm 3 mole '. The experimental value
is 4230 x 10-6 cm 3 mole-', whereas the spin-only formula predicts a value of
3359 x 10-6 cm 3 mole -1.
There is one additional point to be made in this section. If electrons are
delocalized onto the ligands as a result of covalency in the metal ligand bond,
the matrix elements corresponding to orbital angular momentum are reduced
below the value calculated by using the metal-centered operator L. To compensate
for this, one can use kL (where k, the orbital reduction factor, is a constant less
than one) to correct for delocalization of electron density onto the ligand where
it will have reduced orbital angular momentum.
11-4 APPLICATIONS OF SUSCEPTIBILITY MEASUREMENTS
Spin-Orbit Coupling
When equation (11-34) is applied to complexes of the rare earth ions, excellent
agreement between calculated and observed susceptibilities results, as shown for
some trivalent ions in Table 11-4.
TABLE 11-4. Calculated and Experimental Magnetic Moments for Some

Trivalent Rare Earth Ions
Element Config. Term pte(calc) p.,(exp)
1
Ce" 4f 2Fs/2 2.54 2.4
Pr 3 4f 2 3H4 3.58 3.5
Nd" 4f 3 4J9/2 3.62 3.5
Pm 3" 4f 4 514 2.68
Sm3" 4f 5 6
H5 /2 0.84 1.5
Eu3 4f6 7F0 0 3.4
Gd 3 4f 7 S 7.94 8.0
Tb3 4f8 7
F6 9.72 9.5
Dy 3 4f 9 6
H15 2 10.63
Ho3f+ 41 o -8 10.60 10.4
Er 3 4f 1 415/2 9.59 9.5
Tm" 4f12 3H 7.57 7.3
2
Yb 4f13 F 7/2 4.54 4.5
TABLE 11-5. Calculated and Observed Magnetic Moments for Complexes

of the 3d Ions
(Calculated results are presented using equation (11-34) and the spin-only
formula.)
Ion Config. Term g[J(J + 1)]11 2[S(S + 1)]1I2 len(exp)
Ti +, v4 3di 2D3/ 1.55 1.73 1.7-1.8

V3 3d2 3F
2
1.63 2.83 2.6-2.8
Cr 3 , V2 + 3d 3 4
F3 /2 0.77 3.87 ~3.8
Mn3+, Cr 2 3d4 5DO 0 4.90 ~4.9
Fe3 , Mn 2 t 3d5 6S5/2 5.92 5.92 ~5.9
Fe 2 + 3d6 5
D4 6.70 4.90 5.1-5.5
Co 2 3d7 4
F9/2 6.63 3.87 4.1-5.2
Ni 2 + 3d8 3
F4 5.59 2.83 2.8-4.0
Cu 2 9 2
3d D5/2 3.55 1.73 1.7-2.2
This excellent agreement is obtained because the crystal field from the ligands
does not effectively quench the orbital angular momentum of the electrons in the
inner 4f orbitals. A very much different result is obtained with the 3d transition
series where, as can be seen in Table 11-5, the spin-only formula comes much
closer to predicting the observed results.
Thus, in many of the complexes, the orbital contribution is largely quenched
by the crystal field. There is a very simple model that enables one to predict
when the orbital moment will not be completely quenched. If an electron can
occupy degenerate orbitals that permit circulation of the electron about an axis,
orbital angular momentum can result.* There must not be an electron of the
same spin in the orbital into which the electron must move.
In an octahedral d' complex, for example, the electron can occupy d., and
dy2 to circulate about the z-axis, and the complex possesses orbital angular
momentum. In an octahedral, d3 complex, there are electrons with the same spin
quantum number in both d,, and d,, so this ion does not have orbital angular
* This rule is often expressed by requiring that a rotation axis exists that enables one to rotate
one of the degenerate orbitals containing the electron into another vacant one.
momentum. Using this crude model, we would predict that the following
octahedral complexes would have effectively all of the orbital contribution to the
moment quenched:
High-Spin:
3 3 1 3 2 6 2 6 3
t2|, t2g e , t2| e|, t6 e2, t2g e
Low-Spin:
t2g6 1
t2g6 and
The one electron in eq could only occupy d.2 and d2 _,2 so circulation about an
axis cannot result. If an E state in an appropriate ligand field consisted of an
electron in dY and d2 _,2 orbitals, an orbital contribution would be expected.
For tetrahedral complexes, for which high spin complexes result, the orbital
contribution is quenched in el, 12 2 3 3 3 43
e , e t2 , e t 2 , and e't 2 .
Many molecules with A 2g and Eg ground states have moments that differ
from the spin-only value. This variation results from two sources: (1) mixing in
of excited states that have some contributions from spin-orbit coupling, and (2)
second-order Zeeman effects (temperature-independent paramagnetism). For
example:
[eff(A 2g) = p(spin only) 1 1 (11 36)

' I ODqJ lODq
The last term is the temperature independent paramagnetism which is field

induced. The 4/1ODq term arises from the mixing in of an excited state via
spin-orbit coupling. In lower symmetry complexes, the states are split and more
mixing becomes possible. For an Eg ground state, the moment from mixing in
an excited state is given by:
212 4Np2
Petf(E) = p(spin only) 1 10Dq 10Dq (11 37)
As we can see from the above discussion, the magnetic moments of transition
metal ion complexes are often quite characteristic of the electronic ground state
and structure of the complex. There have been many reported examples of this
kind of application. A few nickel(II) and cobalt(II) complexes will be discussed
here to illustrate this application.
In an octahedral field, nickel(II) has an orbitally nondegenerate ground state,
3
A 2 (t 2g62), and no contribution from spin-orbit coupling is expected. The
measured moments are in the range from 2.8 to 3.3 BM, very close to the spin-only
value of 2.83 BM. Values for octahedral complexes slightly above the spin-only
value arise from slight mixing with a multiplet excited state in which spin-orbit
coupling is appreciable. Tetrahedral nickel(II) has a 3 T1 ground state that is
essentially (eg4 t2q 4 ), and a large orbital contribution to the moment is expected.
As a result, even though both octahedral and tetrahedral nickel(II) complexes
contain two unpaired electrons, tetrahedral complexes have magnetic moments
around 4 BM compared to 3.3 BM or less for octahedral complexes. Experimen-
tally, it is found(',") that NiCl 4 2 -, Ni(HMPA)4 "' (104) (HMPA = hexamethyl-

phosphoramide), and(""b) NiX2 -2(CH,) 3AsO (X = halogen) have moments in
excess of 4 BM. An inverse relationship exists between the magnitude of the
moment and the distortion of nickel(II) complexes from tetrahedral symmetry.
The complex NiX2 - 2(CH,) 3 P, which is known"' to be seriously distorted, has
a moment of about 3 BM. The structures of NiX 2 - 2HMPA and CoX 2 - 2HMPA
(where X = Cl-, Br , I-, NO 3 ~) were deduced from combined magnetic and
spectroscopic studies. (2) This work provides a good illustration of the use of
these techniques.
In octahedral cobalt(II) complexes, the ground state is 4 T,,, and a large
orbital contribution to the moment is expected. Mixing in of an excited state
lowers the moment somewhat but a value in excess of 5 BM is usually found
[pu(spin-only) = 3.87 BM]. The ground state for tetrahedral cobalt(II) complexes
is 'A 2 and a low moment approaching the spin-only value might be expected.
However, an excited magnetic state is comparatively low in energy in the
tetrahedral complexes and can be mixed with the ground state. Moments in the
range from 4 to 5 BM have been predicted(1 3 1and are found experimentally. An
inverse relationship exists for tetrahedral cobalt(II) 1 3 ) complexes between the
magnitude of the moment of a complex and the value of Dq as predicted by
equation (11-36).
11-5 INTRAMOLECULAR EFFECTS
In our treatment so far, we have assumed that there is no interaction between

the electron spins on the individual metal ions in the solid. Next we wish to
consider molecules containing more than one metal ion with unpaired spins.
Consider the planar M 2 L4 X 2 system:
L M -X -L
L ''X WL
The influence of one metal, i, on the other, j, can be treated as arising from
pairwise interactions of metal d orbitals on centers i and j leading dimeric m.o.'s,
fr and Oj. When two d orbitals, each containing one electron, are orthogonal,
the two m.o.'s are degenerate, a triplet state arises, and the coupling is said to
be ferromagnetic. When the two a.o.'s undergo a bonding interaction, Orand
differ in energy. If the energy difference is greater than the pairing energy, a singlet
state arises and the metals are said to be antiferromagnetically coupled. The
Hamiltonian for the interaction between a pair of spins on atoms i and j is given
in equation (11-38)
= JS -$ (11-38)
In the literature, one finds equation (11-38) written with -J, 2J, or -2J
instead of J. With the convention in equation (11-38), J is positive for an
antiferromagnetic interaction and negative for a ferromagnetic system.
11-5 Intramolecular Effects 487
Substituting the energies for S, = S- 2 from Fig. 10-8 into equation
(11-32) produces
2(g#3) 2 (-J) 2g2fl 2

Z kT 4kT kT
N ( -J\ (+3J\ (J
p4k T)+ 4kT) 3epkT)
Rearranging, we find
2Ng 2 3 2 1
3kT 1+ exp (J/kT)
3
Since Nfl 2 3k is about 1/8 we obtain when g = 2
1 1
x a -(11-39)
T exp (J/kT)
1+
3
From this expression we find that y approaches zero as T approaches infinity

and becomes small as T becomes small. As a result ) must have a maximum,
which is given by setting 8 In x/BT equal to zero. From this we find
J 8
kTN 8 (11-40)
i.e., Z has a maximum value at

5J
TN 2k
When J < kT or when T > TN, the susceptibility follows the Curie-Weiss
law, i.e.,
3 1
X= ±O (11-41)
where 0= J/4k. We also see that as the quantity J approaches positive infinity,
y approaches zero.
Magnetic susceptibility measurements have provided information('4 ) con-
cerning the influence of electronic and structural change on the magnitude of J,
the energy separation of the triplet and singlet states: A molecular orbital
analysis" 5 ) provides a model to explain these influences on the magnitude of J.
To a first approximation, the energy separation of the singlet Es and triplet Er
states is given by
J= ET - Es = -2Kab + a--
Haa - ab(142
where H refers to a two-electron Coulomb integral, K an exchange integral, and

91 and E2 the m.o. (i, and 02) energies. When the two m.o.'s are degenerate,
E1 = E2, J = - 2K,, and a ferromagnetic system results. When the energies differ,
the second term dominates and the splitting leads to a singlet ground state. The
value of J, which corresponds to the energy difference of the singlet and triplet
states, has contributions from a direct overlap of the metal orbitals and from an
indirect mechanism (referred to as superexchange) in which the orbitals interact
through the intermediacy of lone pairs or nr-electrons of a bridging ligand.("I
Orbital interactions involving orbitals on the bridging atom that can lead to
antiferromagnetic or ferromagnetic interactions are shown in Fig. 11-6(A) and
6
(B), respectively,(1 )
Y III
- VII
FIGURE 11-6 Orbital
overlap and molecular B' Q M'
,
d - orbital
M
d - orbital
orbitals leading to (A) an
antiferromagnetic interaction
and (B) a ferromagnetic
o00 0 ($1) ($3)
Molecular
orbitals of the
B($2)
Lone pair
s-orbital
interaction. M represents the bridged system

same metal with one
electron in its d orbital and B
the bridging group. In (A) B
has a pair of electrons in an
s orbital and in (B) pairs of
electrons in two orthogonal p
orbitals.
JIIM6 M
d - orbital
M
d- orbital Molecular
Blone pair
p-orbitals
~0
orbitals of the
bridged system
In complexes in which the M-B-M angle is 90 , the situation is close to

that in Fig. 11-6(A) with the low energy 2s orbital of chlorine making only small
contributions to the bonding. As the angle changes to 1800, more s character is
employed in the chlorine hybrids used to bridge the two atoms and the situation
described in Fig. 11-6(A) results. Experimentally, larger values of J are found as
the bridge angle increases to 180 . The J values of a large number of di-
p-oxo-bridged copper(II) complexes whose structures are known from x-ray
crystallography have been compiled."') For square planar and square pyramidal
geometries the unpaired electron resides in d2 _ Y orbitals. The J values for the
dihydroxo-bridged complexes vary linearly with the bridge angle being ferromag-
netically coupled when the angle is less than 960. At angles of 105 the J value
is greater than 500 cm-. A general trend of increasing J with increasing angle
is also observed in this system for bridging alkoxo, phenoxo, keto, and N-oxo
groups. Additional studies involving a binuclear Schiff base ligand system shows
that J decreases from 660 cm-1 for the Cu-Cu pair to 150 cm-' for a Cu-Ni
* The combination -Sl(Si + 1) - S2(S2 + 1) is constant for each level of the dimer and is
frequently dropped; i.e., the zero of energy is redefined.
11-6 High Spin-Low Spin Equilibria 489
pair. Ferromagnetic systems arise for Cu-V(IV)O and Cu-Cr(III), which have
electrons in orthogonal d,2 -,2 and de orbitals. In higher symmetry complexes,
with triply degenerate T ground states, the problem becomes more complicated
with the spin Hamiltonian formalism breaking down and more than one
parameter being needed to describe the exchange interaction.
A bimetallic doubly bridged oxo system formed by coordinating
bis(hexafluoroacetylacetonato) copper(II) to the oxygens of a Schiff base copper(II)
complex was found) 8 ) to have a J of only 20 cm-'. The ligand distortions in
the adduct, were determined by x-ray diffraction allowing an assignment of the
location of the dX2 _,2 orbital containing the unpaired electron on the two centers.
Of the two bridging oxo groups only one is common to the two d2 _,2 orbitals.
As a result, this complex does not obey the above correlations for doubly bridged
oxo dimers. The reader is referred to extensive literature"'-21) for additional
studies of the influence of geometry and orbital energies on the values of J.
11-6 HIGH SPIN-LOW SPIN EQUILIBRIA
If one examines the Tanabe and Sugano diagrams for d4 , d', d', and d7 octahedral
complexes, it can be seen that for certain values of Dq/B the ground state changes
from a high spin to a low spin complex. For d4 the 'E9 and 3 Tlg states are
involved, whereas d5 involves 'A 1, and 2 T 2 g, d' involves 'T 2 g and iA 1g, and d7
involves 4 Tig and 2 E9. When the ligand field is such that the two states are close
in energy, the excited state can be thermally populated and the system will consist
of an equilibrium mixture of the two forms. There have been many reported
studies of this type of behavior. 22 al A typical system, which is selected for
discussion because it has been studied 2 2b) in the solid state and in solution,
involves iron(II) complexes of ligands that are Schiff base type condensation
products of tren [N(CH 2CH 2NH 2)3 ] and 2-pyridinecarboxaldehyde. The result-
ing complex is shown in Fig. 11-7, where only one of the pyridine aldimines has
been drawn in for clarity of presentation. The whole series of compounds in
which R, R', and R" are -H or -CH 3 were prepared.
The symbol (I) will be used to abbreviate the complex [Fe(Py)3 tren] 2 +
where R = R' = R" = H; (II) symbolizes [Fe(6MePy)(Py) 2tren] 2 + where
R = R'= H and R" = CH,; (III) symbolizes [Fe(6MePy) 2(Py)tren]2+ where
R = H and R' = R" = CH 3 ; and (IV) symbolizes [Fe(6MePy) 3tren] 2 + where 2 Fe
R = R'= R"= CH 3 - 3
An equilibrium involving the low spin 'Aig(t2, 6 ) and high spin 'T 2 g(t2 g4 eg2 ) N R
4 6 R IR"
states was found both in the solid state and in solution for several of these 5 R'
complexes. Complex (I) is fully low spin at and below room temperature, while
(II) and (III) undergo spin equilibrium both in the solid state and in solution. In FIGURE 11-7 Structural
solution, complex (IV) is essentially high spin at all temperatures above 180 K. formula of
In the solid state, a spin equilibrium exists that is very anion dependent. tris{4-[(6-R)-2-pyridyl]-3-aza-
Thermodynamic data for the interconversion can be determined from the change 3-butenyl} amine iron(li)
in susceptibility with temperature, and are reported to be + 4.6 and + 2.8 kcal complex.
mole-', respectively, for complexes (II) and (III) in solution. An x-ray crystal
structure determination indicated that the methyl substituents on the pyridine
ring interact with the adjacent pyridine moiety. Thus, the ligand field is weakened
in compound (IV) to the extent that a high spin compound results, while
compound (I) is low spin. In the case of compound (IV), the average metal nitrogen
distance is found to decrease by about 0.12 A in going from the high spin to the
low spin complex.
11-7 MEASUREMENT OF MAGNETIC SUSCEPTIBILITIES
In this section, the measurement of bulk magnetic susceptibility will be briefly

covered12 1) in the course of presenting the pertinent references for a more complete
discussion. The Gouy method 2 3b) employs a long, uniform glass tube packed
with the solid material or solution, which is suspended in a homogeneous magnetic
field. The sample is weighed in and out of the field, and the weight difference is
related to the susceptibility and field strength. If a standard of known susceptibility
12
is used, the field strength need not be known. Evans 3c) has reported a very
clever and inexpensive device for routine measurement of the magnetic suscep-
tibility by the Gouy method.
The Faraday method(2 3 a) uses a small amount of sample that is suspended
in an inhomogeneous field such that H(OH/BX) is a constant over the entire
volume of the sample. The method is very sensitive, so small samples can be used
and studies can be made on solutions.
Susceptibilities can also be determined conveniently over a wide temperature
range, down to liquid helium temperatures, with a vibrating sample mag-
netometer.(2 3d) The change in the inductance of a coil upon insertion of a sample
can be related to the sample susceptibility. A mutual inductance bridge has been
described and used for susceptibility determination.(2 3 ") An ultrasensitive,
superconducting quantum magnetometer with a Josephson junction element has
also been described."
When one studies the susceptibility of single crystals, the anisotropy in the
susceptibility can be determined.1 2 1-28) This information has several important
applications, as we shall see in our study of the nmr and epr of transition metal
29 30
ions. The Krishnan critical torque method has been commonly used. , )
31
An nmr method has been reported ) for the measurement of the magnetic
susceptibility of materials in solution. In this method, a solution of the paramag-
netic complex containing an internal standard is added to the inner of two
concentric tubes. A solution of the same inert standard, dissolved in the same
solvent containing the dissolved complex, is placed in the outer of the two
concentric tubes. Two separate nmr lines corresponding to the standard will be
observed, with the line from the paramagnetic solution lying at higher frequency.
The contribution to the absolute chemical shift of a nucleus from the bulk
32
magnetic susceptibility of the sample (Av/v,)x, is given by ) :
m-- )[ - Z( 1 + P ') (11-43)
where Av is the shift in H2, vo the spectrometer frequency, m the mass of solute
per unit volume, ot a demagnetization factor (vide infra), Z and Z, the mass
susceptibility of the solute and solvent, respectively, and p. and p, the densities
of the solvent and solution respectively. x has contributions from a paramagnetic
yZ'and diamagnetic term X as follows:
X Z- Z (t 1-44)
11-8 Superparamagnetism 491
For a spherical sample a = 1/3 and the contribution to the shift from bulk
susceptibility vanishes. For a cylindrical sample whose cylinder axis is perpen-
dicular to the magnetic field direction, t = 1/2. In most superconducting magnets
the magnetic field direction is parallel to the cylinders axis, and Y = 0. Thus,
[(1-) -a] equals - / in the former case and '/3 in the latter.
The shift of the standard in the paramagnetic solution is measured as just
described. The measurement is repeated for an analogous diamagnetic complex
(e.g., the zinc complex). If the concentration of paramagnetic and diamagnetic
complex used is the same, the difference in the two shifts A(Av/v,) is related to
the mass susceptibility x by:
7Av' 71
A -) m - - cJp (11-45)
vo 3
The paramagnetic contribution to the molar susceptibility Z', which is related

to pe, (equation (11-23)), is given by:
fAv' 71
A v) = 1000 M - - P (11-46)
Errors can arise if the reference material undergoes a chemical interaction with
the paramagnetic center. It is best to repeat the experiment with different types
of references.
11-8 SUPERPARAMAGNETISM*
Superparamagnetism is observed in very small particles of transition metals and

their compounds, particularly their oxides. It can be used to characterize fine
dispersions of metal(s), alloy(s), and their oxides and has applications in several
areas including catalysis.
It should be noted that parameters such as the magnetization per cubic
centimeter (M) or magnetization per gram (a) are better parameters than the
susceptibility (y) for describing superparamagnetism and magnetically ordered
ferro- and ferrimagnetic materials. Magnetization (M) is expressed in units of
"gauss"; this unit in reality stands for "gauss per cm 3 ." Quite often "emu/cm 3 "
is employed. Dividing gauss by the density (g/cm 3) gives "gauss cm 3 per g," which
is an awkward unit for the magnetization per gram (a). This situation is often
circumvented by naming it "emu/g." It should be stressed that "emu" is an
abbreviation for electromagnetic system of units in the cgs system and "emu"
itself is not really a unit. The applied magnetic field (H) is expressed in oersteds
(Ge). Furthermore, the magnetic moment in ferro-, ferri-, and superparamagnetism
* This contribution provided by L. N. Mulay is an abbreviated, modified form of his Chapter

42 in "Catalyst Characterization Science," Ed. M. L. Deviney and J. L. Gland, A. C. S. Symposium
Series No. 288, Washington, D.C. (1985). This material is reproduced by permission of the author
and the American Chemical Society.
to the saturation moment (p or p) and not the effective moment (pef)

Hsrefers
3 3 34
Oe which is important in paramagnetism;( , )
G
pefr = n(n + 2)
where n is the number of unpaired spins associated with discrete species such as
ions and free radicals. The saturation moment is given by
G H
ps = n (11-47)
This situation arises because n, the number of spins in a cluster, is very large-
of the order of 10'. A discussion of various systems of units, with factors for
33
conversion to SI units is given by Mulay and Boudreaux ( ) and by Mulay and
FIGURE 11-8 Schema tic (4
diagram of a hysteresi ti Mulay , who discuss many applications to catalysis.
Superparamagnetism is best understood by comparing it with paramagnetism
curve for a typical
ferromagnet showing
and ferromagnetism. Ferromagnetic metals, such as Fe, Co, Ni, and insulators
magnetization (a) as a such as T-FeO 3 and Fe3O4 , show the well-known hysteresis curve (Fig. 11-8),
function of the applied which stems from their domain structure. In their unmagnetized state, the
magnetic field (H). Sat uration unpaired electrons associated with each atom (or structural unit) have a net
magnetization is indic ated by (spontaneous) magnetic moment (it) or magnetization (aj which is the vector
sum of all unpaired electrons in that domain; these are shown by the large arrows
in Fig. 11-9. Saturation is shown by as, the remanence, the remaining magnetic
induction by a,' and the reverse field necessary to bring a, to 0 is given by H~,
the coercive force. Between the domains, a "Bloch wall" is formed, with spins
A2 curled up in a helical fashion. The magnetization present within a ferromagnetic
domain at zero degrees Kelvin and with zero applied field is called the
spontaneous, intrinsic, or technical magnetization ('sp). The subdomain clusters
with varying sizes represent superparamagnetic clusters (Fig. 11-9), with no
interaction between neighbors. These clusters are thermally unstable, that is, their
93 magnetic moments (represented by moment vectors) experience thermal fluctu-
ations with great ease, as is the case with paramagnetic species due to the lack
FIGURE 11-9 Schem atic of long-range ordering.
representation of a materials are
multidomain ferromag neticIt is well known that when ferromagnetic and ferrimagnetic
Curie temperature (T, they change over to paramagnetic
particle in the unmagn etized heated above a critical
etie behavior; thus the hysteresis disappears. In contrast, ideal superparamagnetic
state. Each of the thre e
domains with net mom ients systems, when sufficiently cooled below a critical blocking temperature (TB) will
Pi, p2, P3 remain rand )mized experience a very slow relaxation time. Their net magnetic moment will align
in this state. Circles o f parallel to the applied field (H) and appear to behave as if they had an apparent
varying sizes represe nt the "bulk-like" ferromagnetic behavior. This aspect will result in hysteresis of
"apparent" ferromagnetic behavior. Conversely, above the T, the hysteresis will
subdomain
superparamagnetic cl disappear and the clusters will show a unique curve with no hysteresis (Fig. 1110).
osters.
Each ferromagnetic domain consists of myriads of spins. One can imagine
small clusters within each domain having different volumes (V , Vs. . V). These
subdomain particles, which consist of several thousand spins are the superparamag-
netic clusters. They will have large magnetic moments of several thousand BM
(Bohr magnetons: recall one BM = eh4rmc). When suh superparamagnetic
clusters are well-dispersed on (or "within") a substrate (e.g., i, carbon, zeolites,
etc.), there is no magnetic interaction between their moments. Hence, the
11-8 Superaramagnetism 493
I I FIGURE 11-10 Plot of the

relative magnetization a/a,
as a function of H/T. (A) A
1.0 .r-------------------------------------------------------------- paramagnetic system is
4 p > 106 (C) Ferromagnetism characterized by an effective
per domain moment with a maximum
* , per cluster Bohr Magneton number of
0.8 - 10 per ion and by the
(B)absence of hysteresis. The
very small slope of the
0.6 straight line is exaggerated
for clarity. Paramagnetic
'b saturation occurs at a very
ratio HIT 0 OeK.
0.2-high
0(B) Langevin curve for
fttper ion superparamagnetic clusters.
fA) t a am a~g e.t is.MLarge clusters with relatively
high moments are shown to
reach saturation easily at
lower values of H/T.
Similarly small clusters with
0 20 40 60 80 100 120 140 a net small moment are
shown to reach saturation
HIT (oersted degree with difficulty, at high values
of H/T. (C) Top part of a
hysteresis curve observed in
a ferromagnetic curve.
superparamagnetic clusters behave in the same manner as magnetically dilute

paramagnetic ions. For paramagnetic ions, the susceptibility and the c laiyH)
Pars are good parameters for the interpretation of magnetic results. However, for
the superparamsagnetic clusters with a giant BM number of the per gram ure,
magnetization (a) is a good parameter. The a increases with increasing H, and
with decreasing temperature (T)f.Therefore, it is customary to measure U at
constant field strength and varying temperature, and vice versa. When the relative
magnetization a/a, (where ar is the saturation magnetization) is plotted as a
function of HlT, one obtains an excellent superposition of data points, and there
is no coercive force (HS = 0). Therefore, hysteresis is not observed in Fig. 11s8
in contrast to the bulk ferromagnetic material. Hence, superparamagnetic
behavior maysbe regarded as a phenomenon intermediate between paramagnetism
and ferromagnetism. Note that for pure ferromagnetic materials, without mech-
anical strain, the Hh can be as small as 0.5 oersteds.
Since "superparamagnetism" as the name implies, is similar to paramagnet-
ism, one can apply the classical Langevin function, first derived for the non-
interacting paramagnetic spins, to the non-interacting, ideal S.. clusters. Thus,
a/a, coth (puH/kTf - (kT/p,(H)) (11-48)
Here, e stands for the (giant) magnetic moment of the cluster which replaces the
i for single, isolated paramagnetic spins; k is the Boltzmann constant.
In the above equation, all quantities except p, can be measured. The P, can
be derived for (ideal) superparamagnetic clusters by a curve-fitting procedure (cf
Ref. 3). The gist of obtaining the average volume (v) of a superparamagnetic
cluster lies in the basic definition of o-, which for a superparamagnetic system
can be written as a, = p/v. It should be noted that at the atomic level, the molar
saturation magnetization (Xm) is given by
XM = N/By, emu per mole
here N is the Avogadro's number of species, each carrying a moment p, = n (the

number of unpaired electrons per species), and #l is the Bohr magneton. (A
convenient number, Nfl = 5585).
6
From the low-field (LF) and high-field (HF) approximations(35,3 > of the
Langevin function stated below, one can calculate the vLF for large clusters (which
magnetically saturate easily at low values of H/T) and vHF for small clusters
(which saturate with difficulty at high values of H/T). From the vL F and vHF,3 5the
, 36 )
average volume v of clusters can be estimated by taking the arithmetic mean.
1"
8k auIHT (11-49)
nEI
6k (11-50)
dHF 3 = __ - sI /'(H/T)
nrI,
Here I, is the spontaneous magnetization for the bulk material. For elemental
Fe, I, is 1707 gauss (or oersted) at room temperature. Cluster diameters as small
as 2 nm and ranging up to 20 nm have been deduced from magnetic results for
various catalysts such as Fe, Fe-Co, Fe-Ru, and Ni, supported on high surface
3 4 -36)
area alumina, silica and carbon.(
REFERENCES 1. A. P. Ginsberg and M. E. Lines, Inorg. Chem., 11, 2289 (1972).

2. G. Herzberg, "Atomic Spectra and Atomic Structure," 2nd ed., Dover Publications,
CITED New York 1944, 109.
3. M. Kotani, J. Phys. Soc. Japan, 4, 293 (1949).
4. B. Bleaney and K. W. H. Stevens, Rep. Prog. Phys., 16, 108 (1953).
5. F. E. Mabbs and D. J. Machin, "Magnetism and Transition Metal Complexes," John
Wiley, New York, 1973.
6. a. B. N. Figgis and J. Lewis, "The Magnetic Properties of Transition Metal
Complexes," in "Progress in Inorganic Chemistry," Volume 6, ed. F. A. Cotton,
b. B. N. Figgis and J. Lewis, "The Magnetochemistry of Complex Compounds," in
Modern Coordination Chemistry," eds. J. Lewis and R. G. Wilkins, Interscience,
New York, 1960.
7. R. S. Nyholm, J. Inorg. Nuclear Chem., 8, 401 (1958).

8. N. S. Gill and R. S. Nyholm, J. Chem. Soc., 1959, 3997.
9. R. L. Carlin, J. Chem. Educ., 43, 521 (1966).
10. a. J.T. Donoghue and R. S. Drago, Inorg, Chem., 1,866 (1962) and references therein.
b. F. A. Cotton et al., J. Amer Chem. Soc., 83, 4161 (1961) and references therein.
11. L. M Venanzi, J. Chem. Soc., 1958, 719.
13. F. A. Cotton et. al., J. Chem. Soc., 1960, 1873.
14. G. F. Kokoszka and G. Gordon, Trans. Metal Chem., 5, 181 (1969).
15. P. J. Hay, J. C. Thibeault and R. Hoffmann, J. Am. Chem. Soc. 97, 4884 (1975).
16. R. S. Drago, Coord. Chem. Rev., 32, 97 (1980).
17. D. Gatteschi, "The Coordination Chemistry of Metalloenzymes," ed. I. Bertini, R. S.
Drago, and C. Luchinat, D. Reidel Dordrecht, Holland, 1982.
18. K. A. Leslie, R. S. Drago, G. D. Stucky, D. J. Kitko and J. A. Breese, Inorg Chem.
18, 1885 (1979).
19. a. B. N. Figgis and R. L. Martin, J. Chem. Soc., 1956, 3837.
b. E. Kokot and R. L. Martin, Inorg. Chem., 3, 1306 (1964) and references therein.
c. W. E. Hatfield in "Theory and Applications of Molecular Paramagnetism" ed.
E. A. Boudreaux and C. N. Mulay, Wiley, New York, 1976.
d. D. J. Hodgson, Prog. Inorg. Chem., 19, 173 (1975); "Extended Interactions between
Metal Ions in Transition Metal Complexes, ed." L. V. Interrante, ACS Symposium
Series No. S, ACS, Washington, D.C., 1974.
20. J. A. Bertrand, J. H. Smith and P. G. Eller, Inorg. Chem. 13, 1649 (1974).
21. B. K. Teo, M. B. Hall, R. F. Fenske and L. F. Dahl, J. Organomet-Chem., 70,413 (1979).
22. a. R. L. Martin and A. H. White, Trans. Metal Chem., 4, 113 (1968).
b M. A. Hoselton, L. J. Wilson and R. S. Drago, J. Amer. Chem. Soc., 97,1722 (1975).
23. a. See, for example, L. N. Mulay and 1. L. Mulay, Anal. Chem., 44, 324R (1972).
b. B. Figgis and J. Lewis, "Techniques of Inorganic Chemistry," eds. H. Jonassen and
A. Weissberger, Volume IV, p. 137, Interscience, New York, 1965.
c. D. F. Evans, J. Physics E; Sci. Instr., 7, 247 (1974).
d. S. Foner, Rev. Sci. Instr., 30, 548 (1959).
e. W. L. Pillinger, P. S. Jastram and J. G. Daunt, Rev. Sci. Instr., 29, 159 (1958).
f. J. N. McElearney, G. E. Shankle, R. W. Schwartz and R. L. Carlin, J. Chem. Phys.,
56, 3755 (1972).
24. J. W. Dawson et al., Biochemistry, 11, 461 (1972) and references therein.
25. S. Mitra, Trans. Metal Chem., 7, 183 (1972).
26. M. Gerloch et al., J. Chem. Soc., Dalton, 1972, 1559; ibid., 1972, 980 and references
therein.
27. W. D. Horrocks, Jr. and J. P. Sipe, III, Science, 177, 944 (1972).
28. B. N. Figgis, L. G. B. Wadley and M. Gerloch, J. Chem. Soc., Dalton, 1973, 238.
29. K. S. Krishnan et al., Phil. Trans. Roy. Soc. A, 231, 235 (1933).
30. D. A. Gordon, Rev. Sci. Instr., 29, 929 (1958).
31. a. D. F. Evans, J. Chem. Soc., 1959, 2005.
b. D. Ostfeld and I. A. Cohen, J. Chem. Educ., 49, 829 (1972).
32. W. C. Dickinson, Phys. Rev., 81, 717 (1951).
33. L. N. Mulay and E. A. Boudreaux (Eds.), "Theory and Applications of Molecular
Paramagnetism" and "Theory and Applications of Molecular Diamagnetism," Wiley
Interscience, New York 1976.
34. L. N. Mulay and Indu Mulay, "Static Magnetic Techniques and Applications," Chapter
3 in "Physical Methods of Chemistry," Vol. III B, ed. B. W. Rossiter and J. F. Hamilton,
Wiley Interscience Publishers, New York, 1989.
35. H. Yamamura and L. N. Mulay, J. Appl. Phys., 1979, 50, 7795 (1979).
36. P. W. Selwood, "Chemisorption and Magnetization," Academic Press, New York,
1975.
EXERCISES 1. Consider the following system of energy levels, which could arise from zero-field
splitting of an S = 1 state:
2 g2 # 2 H 2
E(10)) = - - D -
3 3D
1 g 2#2H 2
E(I - 1>) = - D - gfH +
3 6D
2
g / 2H2
1
E(|1>) = - D + g#H +
3 6D
Here D is the zero-field splitting parameter and all other symbols have their usual
meanings. Determine the molar paramagnetic susceptibility of this system.
2. The p' configuration of a free ion has a 2 P ground state. Consider the ion in a magnetic
field directed along the z-axis.
a. What is the degeneracy of this state?
b. Write the wave functions for this state in Dirac notation, i.e., |ML, Ms>.
c. The major perturbation on this state is spin-orbit coupling:
H= AL -S
Consider the major contribution in a strong z field leading to:
Does this give diagonal elements, off-diagonal elements, or both, in the basis set
of part b? Evaluate the energies of these wave functions using ft.
d. The Zeeman Hamiltonian is h = #1 - (L + 25) = fH(L + 25J. Evaluate the

energies of wave functions of part b using this Hamiltonian.
e. Use the energies obtained in parts c and d, and the Van Vleck equation,
to calculate Z.
3. The following represents the approach taken to the magnetic susceptibility data of a
complex of the form shown as follows.
Consider octahedral Fe(II). In many systems, the low spin 'A g state is less than
1000 cm -1 below the high spin 5T 2g state, so that both are appreciably populated at
room temperature. The 'A state is non-degenerate and the 'T is 15-fold degenerate
(why ?). Mathematically, a T state in an octahedral field is equivalent to a P state in
a free ion; i.e., it may be considered to have L = 1 and M, 1, 0, + 1. Consider
the 'A state to be unshifted by all perturbations.
a. Write the 15 basis functions for the 'T state in Dirac notation, using T-P
equivalence. Hint: The wave function having ML = I and Ms = -2 is denoted
|1, -2>. In this formulation, M, and Ms are often referred to as "good quantum
numbers."
Exercises 497
H 3 4
CH 2CH 2-NX N 5
R'
N-CH2 CH2 -NX O

N
+ Fe(II)
R"
R"'.
CH CH
2 -N
2 CQ
H
b. Now, let the complex undergo a trigonal distortion, i.e., one for which the
octahedron's threefold axis is retained. The effect of this distortion is described
phenomenologically by the Hamiltonian
((26 2)
-
where 6 is the trigonal distortion or zero-field splitting parameter and L. is the

z-component of the equivalent orbital angular momentum. As is often the case,
chemical factors influencing the magnitude of 6 are poorly understood, and values
obtained in experiments such as this may help elucidate the factors involved. The
Zeeman Hamitonian for a magnetic field oriented along z is:
H = #H(L + g,S)
Here we are considering only the z-direction (parallel to the threefold axis), so our
final expression will be for zx
1.
The following illustrate the use of the Hamiltonian and wave functions:
2)1, -2> = ML |1, -2> - 1)|1, -2> 1,-2>
5zj|1, -2) = Ms| 1, -2)> -2|1, -~2'
Apply the two terms of the Hamiltonian discussed above to the 'T wave
functions derived in part a. Show that all off-diagonal elements of the 15 x 15
matrix must be zero. (This means that all E(2 ) terms of the Van Vleck equation
will be zero.) Hence, determine the energies of the 15 wave functions. Confirm that,
as the total Hamiltonian has been constructed, the center of gravity is maintained;
i.e., that H =lj, though giving the correct splitting pattern would not maintain
a center of gravity.
c. Let the difference between the 'A and 'T states be parametrized by E. The
energy level diagram should look like this:
5
A
5
T'- E '
E
------------
Oh D3 mag
Use the Van Vleck equation to determine y as a function of E, 6, and T.

In the actual experiment, spin-orbit coupling was included, introducing
off-diagonal elements that complicated the analysis. The experimental X vs. T
relation was computer fitted to the theoretical expression to yield best values for
;, E, and 6. It is interesting to note that a poor fit was obtained unless E was
allowed to vary with temperature.
4. The 300 K molar magnetic susceptibility for a solid sample of
H3 C H
H.C
Cu(hfac) 2-O-N
H
H3 C H
H3 C H
was determined to be =- 186 x 10 6 cm 3 mole-. Calculate the molar paramag-

netic susceptibility by correcting for the diamagnetism of the complex. What is pelf?
How can you explain this Plff?
o0 0 0
hfac - CF
CF3
5. a. Co(N 2 H4 )2 C 2 has a magnetic moment of 3.9 BM. Is hydrazine bidentate? Propose

a structure.
b. How could electronic spectroscopy be employed to support the conclusion in part

a?
6. In which of the following tetrahedral complexes would you expect contributions from
spin-orbit coupling? V 3 *, Cr 3 *, Cu 2 +, Co 2 +, Fe2 +, Mn2+.
7. In which of the following low spin square planar complexes would you expect orbital
contributions? d 2 , d 3 , d 4 , d 5, d6 .
8. Why is Fe2 (CO), (with three bridging and six terminal carbonyls) diamagnetic?
9. Explain why mixing of a D4, component in with a Td ground state lowers the magnetic
moment in nickel(II) complexes.
Exercises 499
3
10. What is the expected magnetic moment for Er" ?
11. In the figure below, effective magnetic moment vs. temperature curves are shown for
two similar tris-bidentate Fe(III) compounds:
S
Fe R)
S N R3
A: R = n-propyl (A)
B: R = isopropyl
4.0-
3.0-
2.0
3.0-
(B)
2.0 -
0 100 200 300

Temperature ('K)
a. Provide an explanation in terms of the electronic structure of an Fe(III) complex

that accounts for the changes in ye,, over the given temperature range for each of
the two complexes. In other words, why are the two curves so different while the
two complexes appear to be so similar?
b. In the above plots the points represent the experimental data and the lines represent
least-squares fits to theoretical equations. With the use of an energy level diagram,
describe what parameters might be used in such a theoretical treatment.
12 Nuclear Magnetic
Resonance of
Paramagnetic Substances
in Solution
IVANO BERTINI AND CLAUDIO LUCHINAT
12-1 INTRODUCTION
In the early days of nmr, there was a widespread belief that one could not detect
the nmr spectrum of a paramagnetic complex because the electron spin moment
was so large that it would cause rapid nuclear relaxation leading to a short
nuclear T, and a broad nmr line. This is the case for complexes of certain metal
ions [e.g., those of Mn(II)], but it is not the case for many others. For instance,
in Fig. 12-1, the proton nmr spectrum' of the paramagnetic complex
Ni(CH 3 NH 2), 2 + is presented (B) and compared to that of CH 3NH 2 (A).
Hinc.- B 0inc.-
CH3
JrH2TMS CH3 TMS NH2
3.081.92 0 79.9 -98.3
V, ppm v, ppm
(A) (B)
FIGURE 12-1 Proton nmr spectra (simulated) of solutions of (A) CH3NH2 and (B)
Ni(CH 3NH2)62+. Note that the relative scales in (A) and (B) differ.
These spectra raise several questions:

1. Why do we see a spectrum for the paramagnetic complex?
2. Why are the observed shifts from tins in the complex so large relative
to those of the uncoordinated ligand? The normal range of proton shifts for most
organic compounds is about 10 to 15 ppm and the shifts in the complex are well
outside this range.
3. Why does the NH proton resonance shift upfield, whereas that for the
CH 3 protons shifts downfield?
We shall answer all of these questions in the course of logically developing
500 this topic and in the process introduce a powerful structural tool.
12-2 Properties of Paramagnetic Compounds 501
12-2 PROPERTIES OF PARAMAGNETIC COMPOUNDS
We shall briefly review and expand upon those properties of paramagnetic

compounds important to understanding their nmr. Unpaired electrons have
magnetic moments whose intensity is given by:
p = g [S(S + 1)]"pB (12-1)
where g, is the free electron g factor, S is the total spin quantum number that is
equal to one-half times the number of unpaired electrons, and pB is the Bohr
magneton*, equal to eh/2m, (9.274096 x 10 - 2 4 J -T -'). Unpaired electrons reside
in molecular orbitals and are delocalized over the entire molecule. The unpaired
electron in one molecular orbital can also spin polarize an electron pair in a
doubly occupied molecular orbital as discussed in accounting for the proton
hyperfine in the epr of the methyl radical. At every point in the molecular frame
there are contributions from the direct (except at a node) and spin polarization
mechanisms. Thus, a nucleus in a molecule, containing one or more unpaired
electrons, may experience spin density at its point in space and/or feel spin density
that is nearby. Referring to atomic orbitals as composing molecular orbitals, net
spin density exists at the nucleus of the atom under investigation only through
its atomic s orbitals since only the wave functions for s orbitals have a finite
value at the nucleus. A node exists at the nucleus in all other type orbitals. In
addition, there are fractional unpaired electrons in orbitals other than s on the
atom of interest and on neighboring atoms. Finally, a large fraction of the unpaired
electrons in the molecule is always located in the paramagnetic metal orbitals.
Magnetic nuclei in a paramagnetic molecule interact with the unpaired
electrons just as a magnetic dipole interacts with another magnetic dipole. The
interaction is formally divided in two parts, one with the spin density at the
resonating nucleus and another with the spin density cloud in the rest of the
molecule. The former is called Fermi contact coupling, whereas the latter is
through-space dipolar coupling.(') The Fermi contact coupling can arise from
either the direct delocalization mechanism or spin polarization. The calaculation
of the dipolar interaction requires evaluation of an integral over all space, along
with a detailed knowledge of the spin density distribution. It is customary to
further divide the dipolar interaction into a so-called metal-centered term and a
ligand-centered term. The former considers the unpaired electrons localized on
the metal center and the interaction is evaluated using a point dipole approxi-
mation. The latter considers unpaired electron density delocalized on the ligand
and is often restricted to one or a few atomic orbitals of the atom in the ligand
whose nucleus is being studied. Indeed, spin density near the resonating nucleus
is always relatively small providing a significant contribution only at small
distances.
In the absence of an external magnetic field, the coupling between a magnetic
nucleus and an electron gives rise to new energy levels. In the case of I= / and
S = 1/2, the coupling partially removes the degeneracy of the four functions shown
in Fig. 12-2(A) leading to Fig. 12-2(B).
* In this chapter we shall use M instead of f#for the Bohr magneton to avoid confusion with
the fl spin state. SI units are used throughout except see is used for seconds.
502 Chapter 12 Nuclear Magnetic Resonance of Paramagnetic Substances in Solution
FIGURE 12-2 Energies of

the spin wave functions for
an S = 1/2, / = 1/2 system in laeaN> IPeAN> (feaN) + aON)
the absence (A) and in the
ceaN> aeI3N> el38N> 10ON A
presence (B) of hyperfine
coupling between the
electron and nuclear spins.
The lower-level and one of (16a>- lae/IN)
the upper-level functions are
(A) (B)
linear combinations of the
non-interacting |Ms, M,>
functions. If the coupling, A,
The application of an external magnetic field* splits the levels according to
is negative, the splitting
the orientation of the electronic or nuclear dipoles with respect to the orientation
would be reversed, with the
single degenerate level of the magnetic field (Fig. 12-3; also see Fig. 9-2). When we consider the
higher in energy. nucleus-electron coupling from the point of view of the nucleus, we should keep
in mind that the orientation of the magnetic dipole of the electron changes very
fast compared to that of the nucleus. Nuclei relax with lifetimes of the order of
seconds whereas electron relaxation leads to lifetimes of 10-' to 10-6 sec for
radicals and 10-7 to 10- " sec for metal ions. 2 ) Therefore, the nucleus sees an
oscillating magnetic dipole corresponding to the electron exchanging between
the positions depicted in Fig. 12-3. To a first approximation the average coupling
ae faN>
FIGURE 12-3 Effect of the

application of a magnetic
field BO of increasing
strength on the spin energy
levels of an S = 1/2, / = '12
system. The labeling of the
spin functions in the high
field limit is shown. The
figure is drawn for a positive
value of A. The electron
transitions are indicated by
arrows.
I eN>
e&aN>
* In this chapter the magnetic induction, BO, will be used instead of the magnetic field, H. The
SI unit for B is the tesla (T) and all equations are presented in this unit. In the discussion, magnetic
induction will be called the magnetic field.
energy is zero because the two electronic orientations provide equal moments
with different signs. However, in a magnetic field, the different populations of the
+ -H-Shorter times E
two m. levels lead to a time average different from zero, as shown in Fig. 12-4.
N m,=l- 2 Longer time
The nuclear coupling with an average electronic spin gives rise to a chemical (13)
shift contribution in the nmr spectrum whose magnitude is proportional to the
nature and extent of the coupling. The magnitude of the shift may be anisotropic FIGURE 12-4 The nucleus,
depending on the orientation of the molecule with respect to the external magnetic N, senses a weighted time
field. In solution, rapid molecular rotation produces an average value of the shift average of the magnetic field
anisotropy. Consequently, this shift is called the isotropic shift or isotropic produced by the unpaired
hyperfine shift and can be determined experimentally by measuring the difference electron in the two energy
in shift of a given atom in a paramagnetic compound and that of the same atom states. The time average of
the electron magnetic field is
in an analogous diamagnetic compound. In the next section on the isotropic
non-zero and along the
shift, we will address the questions concerning the large shifts seen in paramagnetic direction of Bo because the
proton nmr as well as why some shifts are upfield and others downfield. electron spin spends slightly
Electron relaxation provides mechanisms for nuclear relaxation because the more time in the lower level.
nucleus senses a fluctuating magnetic field. In some cases, the effect is moderate
and reasonably sharp nmr resonances result. Both the shifts and nuclear relaxation
times can be analyzed to produce valuable structural and dynamic information.
In other cases, the efficient nuclear relaxation mechanisms lead to broad nmr
lines, often to the point where the spectral line escapes detection. However, even
very broad lines can be exploited. For example, when the protons of water
coordinated to a paramagnetic metal ion are too broad to be detected, the
spectrum can be studied in water as a solvent. When the exchange of the
coordinated water molecules with solvent water is fast compared to the difference
in chemical shift between the bound and free water, a single line results whose
line width is a linear function of the mole fraction of the bound and solvent
water. The smaller the T2 of the bound water proton, the greater the possibility
of detecting line broadening of the average line in dilute aqueous solution. Thus,
information about the structure and dynamics of the systems can be obtained
by understanding the effect of hyperfine coupling on nuclear relaxation.
When the discussion of relaxation is completed, we will understand why the
nmr of some paramagnetic complexes can be detected while others are too broad
to be seen. Finally, examples of the analysis of shift and relaxation data to
chemical problems will be presented along with the use of nmr to investigate the
electronic properties of paramagnetic polymetallic clusters. We will see that
answering the simple questions raised at the start of this chapter provides us
with a powerful tool for studying the structure and dynamics of paramagnetic
compounds.
12-3 CONSIDERATIONS CONCERNING ELECTRON SPIN
So far the discussion has involved pure electron spins. In contrast to the nuclear
spin, the electron moves around the nucleus giving rise to an orbital angular
momentum that is small but different from zero.(') We discussed the coupling of
the spin and angular momentum in Fig. 9-17. The orbital contribution should
be kept in mind when, for the sake of simplicity, we refer to electron spin. The
electron in a magnetic field has two different orientations whose energies are
-('2)geBo and ('/2)gePBB 0 . If, however, the total magnetic moment is larger
or smaller than that of the free electron because of spin orbit effects, the Zeeman
energies will be changed accordingly as modeled in ligand field theory. We account
for these changes by using a g-value that differs from the free electron value, Fig.
12-5. Therefore, we will refer to g. = 2.0023 for the free electron and to g # 2.0023
when spin orbit coupling is taken into consideration. Information about the
actual value of g can be obtained directly or indirectly with epr spectroscopy
(Chapter 13).
E E------- -
FIGURE 12-5 In the

presence of spin orbit
coupling, the electron spin
transition energy differs from g yBB0
gepBBO and depends on the
orientation of the molecule in gpy,
the magnetic field, B0 . The 0 gll#,B,
difference can be expressed - B,
in terms of changes in the
value of the g-factor. In this
example, both the extreme
values gi and g are taken to
be larger than ge
(g1 >g- > g).
The Expectation Value of Sz, ( Sz)
When a bulk sample of a paramagnetic molecule containing one unpaired electron

is placed in a magnetic field, the two m, spin states are not equally populated.
The rate of electron exchange between the two levels is described by the quantity
T2 - The picture is equivalent to that of a single molecule (S = '/2) in a magnetic
field spending more time in one level than the other. The population fraction of
the two spin levels is described by N. (the fraction of spins with m, -- -- '/2 or U)
and N, (the fraction of spins with mn,= + '/2 or #l). One may speak of the
probability, P. and P,, that a single spin has an m, value of -- '/2 or + '/2. The
average value of the spin along the field direction, z, is referred to as the expectation
value of S., (S,). This numerical value is proportional to the modulus of the
vector component along the z-direction of either a single spin or of an assembly
of spins normalized to one spin. A single nuclear spin feels this component of
the electron spin and appropriately the remainder of this section will deal with
the evaluation of an expression for <S,>. Recall that
p = -gpBS (12-2)
The Hamiltonian for the interaction of the electron magnetic moment with the
field was given by
H = -pt -Bo = gPBS - BO (12-3)
For an S = system, applying Boltzmann statistics, we obtain
N
NB = exp (-AE/kT) = exp (-gpB 0 /kT) (12-4)
Solving for the total number of electron spins in the # state, we get
N# = N,,exp ( PBBo (12-5)
Since (gpBB0/kT) 4 1,
exp (±gpBB0 kT)~1 ± (gpB B0 /kT) (12-6)
and equation (12-5) becomes
NP~ N,(1 + gPB 0 /kT) = (N - Nfl)(1 + gpBBO/kT) (12-7)
where N = N, + N, (i.e., the total number of spins). Rearranging this expression

by solving for Nf, we obtain
1 + (gpBBo/kT) N 1 + (gPBB 0 /kT)

N 2 + (g pBBOkT) 2 1 + (gpBB 0 /2kT)
The numerator in parentheses in equation (12-8) is of the form I + nx where

n = 2 and x =gpB/2kT. When x 4 1, we may use the approximation
(1+ x)" 1 + nx to obtain
N (1 + gp~BB0/2kT)2 N
N#~ - (1 + 91B10 /2U) = -(1+ g pB0/2kT) (12-9)
S2 (1 + gpBB0/2kT) 2
Similarly, we can show that
N, ~ (N/2)(1 - gPBB,/2kT) (12-10)
The difference between N, and N. gives the excess spin along the field
direction and is related to the modulus of the resultant vector of the up and
down spin vectors. It represents the net magnetic effect induced in the system's
unpaired electrons by the external field. The magnitude of the expectation value
of S2 is the quantity of interest to us in the nmr experiment. This is given by a

population-weighted average of the +1/2 and -1 2 states:
-gpBB0 (
(+ 1/2 )N +(/2N,
N 4kT
where '/2 and -1/2 are the m,-components along B, of and 1# spins, respectively.
r
There is a direct relationship between <S,> and the bulk magnetic moment
M discussed in Chapter 11. We have seen that the number of excess spins in the
lowest level is twice the expectation value of S. times the total number of electrons
N. This number of excess spins must be multiplied by the component along the
field direction of the magnetic moment of one spin in the lowest state, ('12)gpB,
in order to find the average magnetic moment, that is,
2
M = ('12 )gpB(2N<S>) = Ng IB B 0 4kT
2
(12-12)
This is the Curie law expression, described in Chapter 11 and arrived at by

population weighting of the various levels.
For the general case of more than one spin, <S,> is given by
S
Z S exp (- EM,/kT)
(S ) = S (12-13)
exp (- Es /kT)
M,= -S
Here Em5 is the energy of the state with quantum number Ms and according to
equation (12-1) is equal to gpBMSBO. Thus, equation (12-13) is consistent with
the equation for one electron spin (S = '/2). In equation (12-13), a state with a
particular value of the spin magnetic quantum number S is weighted according
to its equilibrium population. The weighted values are summed over all the energy
levels and divided by the total number of levels to give the expectation value of
S,. If AE << k, then exp (-E/kT) = 1 - E/kT, and after some algebraic manip-
ulations, equation (12-14) results:
(S glPBBOS(S + 1) (12-14)
3kT
For a free electron, the energy separation between two Ms levels is gpB,
with g = 2.0023. The contribution from the orbital magnetic moment (in the spin
Hamiltonian formalism) is accommodated by using values for g that are different
from 2.0023 and different for different orientations of the molecule with respect
to the molecular frame. For a molecule or ion rotating rapidly in solution, the
g value of Equation (12-14) is substituted by an average value of g
(S) =-gaPBBOS(S + 1) (12-15)

3kT
Now that we have an expression for <S,>, we shall proceed to evaluate the
quantitative expression for the nmr contact shift assuming equation (12-15). This
12-4 The Contact Shift 507
holds everytime a single S multiplet is populated. However, when two or three

S multiplets are degenerate (as in the cases of E and T ground states) or close
in energy with respect to k7 the energy separation between MS states is different
from gapBBo, and equation (12-15) becomes a rough approximation. More
thorough calculations are needed for these cases.(4 )
12-4 THE CONTACT SHIFT
The contact shift arises from the presence of unpaired spin density at the resonating
nucleus. It is proportional to the amount of spin density. It is also called the
Fermi contact shift. Fermi provided the first theoretical justification for electron
density at the nucleus in accounting for some details in the electronic spectrum
of the hydrogen atom.(') The contact shift is also referred to as the scalar shift
because it is independent of molecular rotation.
The spin density at the nucleus of an atom in a ligand is a small fraction of
an electron which, in the presence of an external magnetic field, gives rise to a
permanent, time-averaged, additional magnetic field that adds to the applied
magnetic field. Therefore, the nucleus experiences a further shift in resonance
frequency. In other words, the nuclear energy levels are affected by the presence
of spin density at the nucleus as shown in Fig. 12-2(B). If we had a full unpaired
electron at the nucleus, this nucleus would sense a magnetic moment equal to
gpuBKSz> aligned with the external magnetic field, with <S,> given by equation
(12-15). We do not have a full electron at the nucleus whose nmr is studied, but,
in general, only a small fraction of an electron, resulting from direct delocalization
and spin polarization. For simplicity let us consider the case of a single electron
in a molecular orbital T that is directly delocalized on the resonating nucleus
without any spin polarization. The spin density at the nucleus is T(O) 2 where 0
means at 0 distance from the nucleus. The shift from paramagnetism comes from
the magnetic field B' arising from the electron density at the nucleus:
B' = gpB<S>(0) 2
= q 2 pB2 BoS(S + 1)P(0) 2 3kT (12-16)
For more than one unpaired electron, a normalization term (' 2 )S must also be
included. The expression for the shift is:
Av AR 1 S(S +1)
v2 2 IV(0)2 (12-17)
v B, 2S B T
where T(0) 2 is summed over all molecular orbitals containing unpaired electrons.
In quantum mechanical terms, the contact hyperfine term is
H = AI - S. (12-18)
where A, is the contact hyperfine coupling constant and is the same as a discussed
in Chapter 13. Capital A indicates that there may be more than one unpaired
electron in the molecule or ion. A, is defined as:
A, = gNpgptB p() 2
(1219)
3S
If pa, pN (the nuclear Bohr magneton), k, T, 'F, and p, are expressed in SI units,
A, is given in J(A - 2 j 3 T - 2m 2 ). However, it is customary to report the A, value
in terms of A,/(h/rad sec-1) or A,/h (sec-' or Hz). If we substitute the equation
for A, into equation (12-17), we obtain the following expression for the contact
shift:
AB gpBS(S + 1)
0 = A,= (12-20)
BO " NpN k 9NpNBO
We can now see that a very small A, value can give rise to a huge isotropic
shift. For example, an A, value of 12 -27 J gives a contact shift of 40 ppm. The
reader is encouraged to convert the isotropic frequency shifts in Fig. 12-1
(at room temperature) to the equivalent field shifts in tesla. The equality
-AB/Bo = Av/v, where v is the fixed probe frequency, follows directly from the
fact that hv = gNpNbo for nuclear spins.
If one measures the temperature dependence of Av, a plot of Av vs. 1/T
should produce a straight line with a slope proportional to A, for systems
exhibiting Curie law behavior. For systems with an orbitally non-degenerate
ground state, such as octahedral nickel(II) and tetrahedral cobalt(II) complexes,
the application of equation (12-20) is valid; otherwise, equation (12-20) represents
an approximation. Equation (12-20) holds independently of the orientation of
the molecular frame as long as g = g,. It is also valid in solution when an average
value of g is considered.
12-5 THE PSEUDOCONTACT SHIFT
Magnetic moments aligned along the external magnetic field from spin density
at every point all over the molecule (except that at the resonating nucleus) can
also couple with the nuclear magnetic moment. The coupling is dipolar in origin.
The relevant Hamiltonian for the general case of coupled magnetic moment
B, vectors it, = gNpN1 and P2 = gpBS was given in equation (9-23).
N S - 3(S -r)( ' r))

$Hdip = (12-21)
-g1pBgN pN ( 13 - r5
e-
When both I and S are aligned along the external magnetic field (as is the
case when g is isotropic) the energy of interaction is:
FIGURE 12-6 The dipolar E i Il2 (3 cos 2

y - 1) (12-22)
interaction energy between
the electron and nuclear
magnetic dipoles in an
where p, is the nuclear magnetic moment, p2 is the electron magnetic moment,
isotropic situation is a and r is the distance between the origins of the two vectors. The meaning of the
function of the angle y angle 7 is illustrated in Fig. 12-6.
between the electron-nucleus The electron magnetic moment arises from the excess population of the
vector, r, and the external Zeeman levels times the spin density at a given point and r is the distance between
magnetic field, BO. the nucleus and the given point. In order to evaluate the interaction energy over
12-5 The Pseudocontact Shift 509
all space, integration is needed. This dipolar contribution to the electron-nucleus
hyperfine coupling can be measured in the solid state or a frozen solution, with
epr and ENDOR. In solution, where random and isotropic rotation occurs, the
energy described above averages to zero for an isotropic g because the integral
for cos y ranging from 1 to -1 of equation (12-22) is zero. For this averaging
to occur, the rotation rate in radians/sec-', has to be faster than Edip/h. This
condition is easily satisfied in solution even for macromolecules.
Now, let us suppose that p 2 is not isotropic but changes its magnitude as
the orientation of the molecule changes with respect to the external magnetic
field. Under these circumstances, p2 is anisotropic and the integration of equation
(12-22) over cos y is not zero. Thus, an added contribution to the hyperfine
coupling results, that is, an additional magnetic field is sensed by the nucleus
giving rise to a contribution to the chemical shift. The value of p 2 will be
anisotropic when there is an orbital contribution to the magnetic moment which,
by its nature, changes in different directions of the molecular frame (see Fig.
9-17). In this case, g also becomes anisotropic and therefore orientation dependent.
Thus, the same effects that give rise to anisotropy in g also give rise to a shift
contribution. This shift contribution is isotropic because the dipolar effect is
averaged by rotation in solution. Because it is isotropic, like the contact
contribution, it is called a pseudocontact shift and is also referred to in the
literature as an isotropic dipolar shift. The through space nature of the
pseudocontact effect is comparable to the neighbor anisotropic contribution
discussed in Chapter 8, which was seen to be dependent upon differences in XOIA
for different orientations. In a similar fashion, anisotropy for Xpa, leads to the
pseudocontact shift.
The complete evaluation of the pseudocontact shift requires knowledge of
the spin density distribution over the entire molecule. This information is usually
not available. If the unpaired electrons are assumed to be localized only on the
metal ion, a point dipole model leads to equations for evaluating the pseudocon-
tact contribution to the isotropic shift. The general equation in SI units* is(7-8):
Av, 1 3 cos2 0i _ I
v - 41 (1/aNA)z -[ 12(Xxx + x,,)] os'
1 - sin 2 6 cos24i (12-23)

4n(' 2 N)(XXX -3,) s 1-3
rZ
Here, NA is Avogadro's constant, Z., Y, and X,2 are the susceptibility N
components, and the angles 0 and 4 are defined in Fig. 12-7. The coordinates
x, y, and z are defined by the principal directions of the magnetic susceptibility
tensor. r
For cases where the magnetic susceptibility along a given coordinate k can
be approximated by equation (12-24), equation (12-23) can then be rewritten in
terms of g values.
Xkk = YOkk 2B S(S± 1) (12-24)

3kT
FIGURE 12-7 Definition of
the quantities in equation
* Note that x(SI) = 4r x 10- X (cgs emu). (12-23).
The anisotropy in the g tensor is more easily obtained than that in Z, especially
for S = %/2 systems.
Rewriting equation (12-23) in terms of Ag for an axially symmetric system,
we obtain the following equation for the pseudocontact contribution:
Avi(pC) _ 11B2S(S± 1) 3cos2 O 2 2

v 4i 9kT (g g ) (1225)
An example of how these terms are defined for a typical molecule with local axial
symmetry is given in Fig. 12-8.
FIGURE 12-8 Definition of terms in

equation (12-25). The parallel and
perpendicular axes are shown. The 'i
radius vector from the metal center to

the nucleus being investigated is labeled P
r,, and the angle that it makes with the 0 g11 axis
highest-fold rotation axis is 0,. For a
proton shift, r, would be drawn to the
hydrogen of the phenyl ring.
In view of the simplicity of equation (12-25), many studies have been

performed on S = 2 systems. Equation (12-25) is valid when the nucleus studied
feels the unpaired electron as a point dipole (i.e., when delocalization can be
neglected and the nucleus is relatively far from the metal).(') The significant
delocalization of unpaired electrons, at least within the coordination cage, may
prevent the quantitative use of equations (12-23) and (12-25) for donor atom
nuclei or for nuclei close to the donor atoms."')
The foregoing discussion should be made more specific by considering
different types of nuclei. Protons and deuterons only have spin density in an
s orbital and since this spin density is isotropic, a pseudocontact shift does not
result. If the protons are attached to a carbon or nitrogen atom, for example,
bearing spin density in a p, orbital, a pseudocontact shift can arise at the proton
that is then called ligand centered. In this case, ligand-centered effects cannot be
neglected. Combined information on the carbon or nitrogen and an attached
hydrogen atom is needed for a complete treatment. It is also sometimes useful
to obtain independent information from the relaxation parameters (vide infra).
The pseudocontact shift equations are particularly useful for nuclei in metallo-
proteins which are within - 10 A of the metal but are removed from the metal
by many sigma bonds. In this case, the spin density and the contact shift is zero
as is spin density on neighboring atoms.
12-6 LANTHANIDES
Lanthanides are characterized by large spin-orbit effects so that both the spin
and orbital angular momenta are sizable and strongly coupled. This means that
for each electron, the various projections of the orbital angular momentum m,
will interact differently with the various projections of the spin angular momen-
tum, m,, giving a total angular momentum component equal to m, + ms = m.
The m, values identify a j multiplet, just as m, and M, were related to the I and
12-6 Lanthanides 511
TABLE 12-1. Some Electronic Properties of Lanthanide Ions
Ion Configuration 2S+ 'Li of Ground State (multiplicity) gi (Sj/' <S> p

(PpM)
2
Ce3* 4fi Fs/2(6) 6 1.07 0.98 +1.6
2 3
Pr3+ 4f H 4 (9) 4/5 3.20 2.97 +2.7
Nd3+ 4f 3
419 2(10) 8/1 4.91 4.49 +1.05
Pm 3 + 4f 4 51 (9) 3/ 4.80 4.01 -0.6
3 5 6
Sm * 4f H5 /2(6) 2/3 1.79 0.06 +0.17
Eu 3 + (Sm
2
+) 4f 6 7
F(1) - -10.68 -1.01
3
Gd * (Eu +)
2
4f 7 8S 7/2(8) 2 -31.5 -31.50 0
Tb3+ 4f 8 7
F6013) 32 -31.5 -31.82 +20.7
3
Dy * 4f 9 6
H 15i2(16) 4/3 -28.30 -28.54 +23.8
3 0 5/4 -22.50 -22.63 +9.4
Ho* 4f 1 5(17)
Er3+ 4f"1 4115(16) 6 -15.30 -15.37 -7.7

2 3
Tm3+ 4f1 H03) 7/' -8.17 -8.21 -12.7
2
Yb 3
+ 4f 13 F 2(8) 8/, -2.57 -2.59 -5.2
'Calculated from equation (12-27): the formula does not hold for f6 ions.
'Calculated according to equation (12-26).
1
'Calculated by inclusion of the excited states.( 27
2
"Dipolar shift predicted from Equation (12-28) for r = 300 pm, 3 cos 0 1 1, T = 300 K, axial symmetry, and D- values for each
lanthanide estimated from reference 11.
1
'Including contributions from excited J manifolds." For f" ions, the contribution of the ground state manifold is zero.
s manifolds. When electrons are allowed to interact, then J terms are obtained.
The ground states for lanthanides in the oxidation state +3 are reported in Table
12-1."(1i,2 They are characterized by S, L, and J. For example, cerium(III) has
one electron in an f-orbital; S = '/2 combines with L = 3 to give both J = 3 +
/2 = /2 and 3 - 1/2 = '/2. The latter term is the ground state. Although it is not
completely correct, we will consider only the ground state in our analysis. This
is an even poorer approximation for actinides, which have larger spin orbit
coupling and excited states of different J multiplicity closer in energy than the
lanthanides.
The f orbitals are often assumed not to be involved in covalent bonding
with the donor atoms. Little electron delocalization on the ligand molecules is
assumed to result and the contact contribution is often neglected. The validity
of this assumption depends on the number and type of bonds between the metal
ion and the resonating nucleus. It is not justified for nuclei directly coordinated
to the metal ion. We should next consider the spin density at the resonating
nucleus and <S,>, which is now labeled <S>,. It is generally assumed that
crystalline field effects are much less than kT and completely neglected in the
evaluation of S,>,12) :
(S.),= gj(gj - 1) J(J + 1) BBO

(12-26)
3kT
where
J(J + 1) - L(L + 1) + S(S + 1)
gj + - (12-27)
2J(J + 1)
The numerical values are reported in Table 12-1.

The values of <S,> are shown in Fig. 12-9.(1 112 Note that whereas <S,>
values for unpaired electrons in d orbitals are negative, they can be either negative
or positive in lanthanides. A positive value results when the orbital contribution
to the energy level is larger than and different in sign from the spin contribution.
40 40
30- -30
FIGURE 12-9 Calculated
patterns of (o) pseudocontact 20 - -20
shifts(") and (m) - < Sz >
values 12 ) for lanthanide ions. 10- / 10
The shifts are for a nucleus -
3A from the metal with 3 0 .- 0
cos 2 0 1equal to 1 at 300 K. ---
-10- -10
-20 1 -20
La Ce Pr Nd PmSm Eu Gd Tb Dy Er HoTm Vb Lu
Due to the small contact terms, the point dipole approximation for the
pseudocontact contribution is a better approximation for lanthanides that for
first row transition metal ions. The ligand field splitting causes the magnetic
anisotropy. The D., D, and D, parameters take the ligand field splitting into
account. As long as D is larger than the Zeeman energy and smaller than kT,
the expression for the pseudocontact term is:(" 1)
2 2
'6k~ipe) o 9 PB j(j + 1)(2J -1)(2J ± 3)
2
v 4n 60(kT)
D(3 cos 2 O8- 1) + (D, - D,) sin 2 O6cos 2i (12-28)

ri3
The shifts at 300 K for a nucleus i at 3A with 3 cos 2 0 1 = 1 and Q = 0 are

reported in Fig. 12-9.
12-7 FACTORING THE CONTACT AND PSEUDOCONTACT

SHIFTS
In principle, the observed isotropic hyperfine shifts contain contact and

pseudocontact contributions. The latter can be further divided into a metal-
centered and ligand-centered contributions. In lanthanides, the metal-centered
pseudocontact term prevails at least for nuclei a few angstroms away from the
metal. Furthermore, since equations are available for both contact and pseudocon-
tact terms, attempts have been made to factor out the two contributions.( 12 '3 )
If we consider a series of complexes differing in the metal ion, with the same
nucleus and same contact hyperfine constant, the factorization is possible." 3 1, 4 )
12-7 Factoring the Contact and Pseudocontact Shifts 513
TABLE 12-2. Contributions to the Hyperfine Shift of Carbon Atoms in a

Porphyrin Ring of a Low Spin Iron(III) Complex" 7
Contribution et Pyrrole "C" p Pyrrole "C" Meso "C Pyrrole "CH 3
(Av/vo)"o 101.5 36.0 73.2 46.8

(Av/v,)d 31 11 22 5.0
(Av/vO) 36 78 -10
(Av/v O)'o 34 -53 61 47
"The values for a and # carbons refer to TPP; those for meso and pyrrole CH 3 carbons refer to
etioporphyrin.
Let us consider two different metal complexes, i and j, of similar geometry. The
measured shifts for a given proton are given for axial symmetry by
=Av
A(S> + D -G
( =
A(S) + Dj -G
Here, the unknowns are A and G for each nucleus, whereas <S,> and D.
depend only on the lanthanide, and reasonable estimates of their values are
available.(11 1 2 , 5 ) By using these estimates for <S,> and D., we are left with two
equations and two unknowns. If more than two metal complexes are studied,
further equations can be used without introducing more unknowns. More
complicated methods that do not involve the assumption of axial symmetry have
been proposed.(' 6 )
For first row transition metal ions factoring is more complex. Ligand-
centered effects complicate the interpretation of nuclei other than protons and
deuterons. In Table 12-2, the relative contributions to the hyperfine shifts of
carbon atoms in a porphyrin ring containing low spin iron(III) are shown. The
contact and the two pseudocontact contributions are similar in magnitude. For
protons and deuterons, the metal-centered pseudocontact contribution can be of
the same magnitude as the contact term.
For orbitally non-degenerate systems with excited states far removed in
energy, the orbital contribution is small and so is the magnetic anisotropy.
Accordingly, the pseudocontact shifts are small. Examples include octahedral
nickel(II) and tetrahedral cobalt(II). Some anisotropy in the magnetic suscepti-
bility arises from zero field splitting as in the case of high spin iron(III)
complexes. 1 ) Orbitally quasi-degenerate ground states provide large magnetic
anisotropies and large pseudocontact contributions (e.g., octahedral cobalt(II)("
and pseudotetrahedral nickel(II) 20 ) systems). The knowledge of the magnetic
susceptibility tensor provides a direct tool for the evaluation of pseudocontact
contributions. Some examples are reported in Table 12-3 where the magnetic
susceptibility information obtained on single crystals is transferred to the molecule
in solution.
When equation (12-24) is valid (e.g., S = % systems), single-crystal epr studies
provide g values and an estimate of the pseudocontact shifts through equation
(12-25).(27)
TABLE 12-3. Isotropic Shifts and Pseudocontact Con-

tributions from Known Magnetic Susceptibility Tensors
(A) 1H and 3 C Isotropic Shifts (ppm) of the Pyridine
Signals in the bis-Pyridine Adduct of Cobalt(II) bis-
Acetylacetonate Using Single Crystal Magnetic Sus-
1 22
ceptibility Data(2 , )
Atom AV)
2-H+ 32.9 -39.5

fl-H +5.0 -18.1
y-H -9.4 -15.6
2-C -199 -92.5
f#-C +229 -35.7
Y-C -73.8 -28.3
(B) The Aromatic Protons in Nickel(II) Bis-

Salicyladiminate Using Single Crystal Magnetic Sus-
20 23 4
ceptibility Data( , ,2 )
A)SO (Av\ dip
Proton Structure
CH 3 CH 3
3 CH -23.7 +1.4
4 N +19.0 -2.3
5 'Ni/2 -23.7 -4.5
6 7 +3.7 -8.4
(C) Phenyl Protons in Dichlorobis(triphenylphos-

(25 2 6
phine)Cobalt(II) and Nickel(lI) , )
Metal/Proton )s A)dip
(V0 V0
ortho -10.35 -2.13

Co meta +8.08 -0.54
para -11.51 -0.53
ortho -6.38 +0.56
Ni meta +8.03 +0.14
para -11.16 +0.14
12-8 THE CONTACT SHIFT AND SPIN DENSITY
As mentioned, if the contact contribution is dominating the isotropic shifts or if

we have factored out the pseudocontact shift, a measure of the contact hyperfine
coupling constant is available. The contact contribution is then related to the
actual spin density on the resonating nucleus through equation (12-19). Spin
density arises from a direct delocalization mechanism and from spin polarization
of an orbital occupied by two electrons. In principle, both contributions can be
calculated through a molecular orbital treatment. If K is the hyperfine coupling
due to a full electron on an atom (listed inside the back cover), then the direct
delocalization contribution is obtained from equation (12-29)
A, KP
(12-29)
h 2S
where Ad is the contribution from direct delocalization and p/2S is the spin density
28
normalized to one electron.( )
In the 1970s Drago et al. pioneered the calculations of direct delocalization
on hydrogen nuclei through extended Huckel calculations.( 2 9 ) Unpaired spin
density delocalizes through the highest occupied orbital and spin density is
evaluated by calculating (e) at the hydrogen nucleus. The orientation of the
electron in the magnetic field is such that the m, ground state is -1/2. Direct
delocalization from the metal implies that the same sign is maintained; then <S,>
is negative and from equation (12-20) the shift in Hz is positive or downfield. If
we maintain the sign of <S,> negative, we have to change the sign of the hyperfine
coupling constant when spin polarization leads to spin density with an m, = 1/2
ground state. The aliphatic protons of octahedral amino complexes of nickel(II)
and cobalt(II) experience direct delocalization and shift downfield. The amino
protons, on the contrary, experience an upfield (negative figures in Hz) shift
because the large spin density on the nitrogen atom makes spin polarization
dominant at the proton. In these cases, the unpaired electrons are in the e. metal
orbitals, which have the correct symmetry to form a a bond to the nitrogen lone
pair. When the unpaired electrons are present only in t2g metal orbitals in
octahedral complexes, they cannot contribute to the direct 0-delocalization. Spin
polarization of the a molecular orbitals is needed to convey unpaired spin density
onto the ligand. This is the reason that early transition metal ions give rise to
opposite shift patterns compared to those on the right hand side of the series.
In the metal-ligand molecular orbital scheme, the antibonding orbital containing
the unpaired electron (Fig. 12-10) is partly constituted of the ligand highest
occupied molecular orbital (HOMO) when the metal unpaired electron is in an ea
eg orbital. (Usually the HOMO contains the lone pair.) The coefficients of the
atomic orbital in that molecular orbital determine the spin density at nuclei. Spin
delocalization through the highest occupied molecular orbital decays rapidly with FIGURE 12-10 Unpaired
the increasing number of a bonds from the metal to the resonating nucleus. Some spin density transferred onto
examples are reported in Table 12-4. The treatment for carbon nuclei differs the ligand by sigma bonding
because direct spin delocalization on a p orbital induces spin density contributions between the eg metal orbital
at the carbon nucleus from is and 2s orbitals through spin polarization. This containing one electron and
complicates the problem.( )
30 the ligand donor lone pair UL-
TABLE 12-4. Estimated A,/h Values (MHz) for 'H, "3C, and 4 N Nuclei of
30
n-Ethyl Amine Coordinated to bis-Acetylacetonate Nickel(ll)( )
3 13
NNH 2 CH 2 C'H 2 CH 3 C'H 3
14.4 -2.64 -0.70 1.02 1.80 0.22
Spin polarization effects are, in general, hard to handle with theoretical tools.
Some results have been obtained in the case of aromatic moieties since independent
tests come from the investigation of radicals through epr. Assume some spin
density is present in a p. orbital of a carbon atom that is hybridized sp2 . To first
order, it does not contribute to contact contributions on either the carbon or
the attached hydrogen nuclei. However, spin polarization (Fig. 12-11) induces
spin density at the carbon nucleus through the Is and 2s orbitals as well as to
the nuclei of attached atoms through the a covalent bond. An attached proton
therefore experiences spin density at the nucleus. The larger the spin density on
the p. orbital, the larger the spin density at the hydrogen atom. The propor-
tionality constant for this relationship is constant for various systems as long as
the nature of the orbitals involved is the same.
A Qpc" (12-30)
H C
= 2S
FIGURE 12-11 Positive p,
where A, is the hyperfine constant due to spin polarization, pc" is the n spin
spin density on an sp2
density on the carbon atom, and Q is the proportionality constant, which is about
carbon induces unbalancing
of the spin density on the -70 MHz.13 " The negative sign is due to the fact that spin polarization induces
doubly occupied C-H spin density on the hydrogen atom of different sign than the spin density in the
bonding a orbital. The sign of p, orbital. If the carbon bearing the spin density in the p. orbital is attached to
the induced spin density is a CH 3 moiety, the Is orbital of the CH 3 protons can directly overlap with the
positive in the outer region p. orbital and then directly transfer spin density with the same sign onto the
and negative in the inner hydrogen nuclei. The proportionality constant Q(C-CH 3) is now positive and
region of the a orbital. is predicted to be around 75 MHz when rotation around the C-C axis is fast
enough to provide a space-averaged hyperfine coupling constant.(3 2 ) Experimen-
tally, the value of Q is found to vary considerably from one system to another.
Furthermore, delocalization through the 7r molecular orbital leads to large values
of contact shifts for nuclei at a large distance from the paramagnetic center.
In the case of nuclei of the second row of the periodic table, there are several
contributions to spin density at the nucleus. As previously mentioned, the unpaired
spin on a p. orbital induces spin density through Is and 2s orbitals of the same
atom. Furthermore, the p. spin density of the nearby atom contributes, via spin
polarization of the a-bond electrons, to the spin density at the resonating nucleus.
For an sp 2 carbon atom, the contributions to spin density through the 2s orbital
are given by(3 3 ):
A Q + 3 QC
(12-31)
i=1 i=
where Xi are the three atoms bound to the carbon under consideration, Qx. is
a constant describing the polarization effect of the spin density in the p, carbon
on the two electrons of the C-Xi bond, and Qxic is the constant relative to the
polarization effect of the spin density in the p. orbital of Xi on the same C-X
bond. Some of these Q values are reported(3-3) in Table 12-5. To this value
of A,, a contribution from the polarization of the Is orbital should be added.
const pe" (12-32)
where the const is -35.5 MHz.(3 3 )

TABLE 12-5. Q Values (MHz) for Fragments Involved in 7 Electron Spin

Delocalization Systems- 3 3 -35)
CH Oc H F F
0CH CH CC CC OCCH 3 OCF O[C
+ 54.6 -65.8 +40.3 -39.0

-75.6 +53.2
+14+ + 115
-410 +2370
Spin density on p. orbitals are present when there is 7z spin delocalization

from the metal.(3 6) For example, vanadium(III) is d2 with two unpaired electrons
in t 2 , orbitals in octahedral complexes. These orbitals have the correct symmetry
to overlap with 7r ligand molecular orbitals. It should be noted, however, that
the 7r ligand molecular orbitals can be either empty (Fig. 12-12(A)) or full (Fig.
12-12(B)). In the latter case, the description is similar to that of the a bonds. The
FIGURE 12-12 Unpaired

spin density can be
transferred onto the ligand
by 2-bonding between a t2g
t2t 25
metal orbital containing one
I E
unpaired electron and an
empty (A) or full (B) ligand 7T
orbital.
unpaired electron resides in an orbital that is partly constituted of the ligand

highest occupied molecular orbital. This mechanism is called ligand-to-metal
charge transfer. Electron delocalization from t2g directly into the ligand 71 system
occurs when the 21 orbital is unoccupied. 7r delocalization is improbable in ligands
having only o- bonds. The lowest unoccupied molecular orbital of the ligand is
far removed in energy and we should also consider whether the d electron has
the correct symmetry to interact with the ligand orbital or not. In the latter case,
spin density on the ligand occurs through spin polarization.
All the ligands of the type
CH-CH
up HC C--- X
CH-CH
down up
where the X donor group can be -NR 2 , -N=CHR, or 0-,- experience

unpaired spin density in the 71 system that arises from non-orthogonality between
the coordination bond and the n1system. The ortho signals are shifted upfield
(i.e., the spin density is negative (excess of spin = /) on the hydrogen atom and
TABLE 12-6. 'H nmr Contact Shift Patterns (ppm) on Some Aromatic Ligands
L 2 3 4 5 6 Ref. No.
CH, H a -4.8 1.7 -5.2 - - 37
b -12.8 10.7 -17.1 - - 40,41

'C
c -8.2 8.6 -11.0 - 25,42
d -6.9 7.9 -11.3 -

NH,
CH2 e -1.5 1.4 -1.6 - 51
N f 72.4 23.1 6.2 - 21, 43, 52
N g 87.7 26.6 8.0

N
CH, h 55.0 12.0(CH 3) 52.0 71.0 - 47
R
0 i - -25.1 21.3 -19.2 12.1 20, 24, 39

"In Ni(2,4 pentanedionate) 2 L2 .
b1
In NiL 62*
'In CoL2 Cl 2 -
dIn NiL C12-
eIn NiL 6 .
f1n Co(2,4 pentanedionate) 2 L2.
gin Ni(2,4 pentanedionate) 2L 2 -
hin CoL *.
i in NiL 2 (R = i-propyl).
positive on the p. orbital of the carbon atom). The para proton experiences a
similar shift both in magnitude and sign. The meta protons experience downfield
shifts. This alternation of spin density on six-membered aromatic rings is
characteristic of 7r spin delocalization and arises from a node in the 7T system
leading to spin polarization of the meta carbon.
The same pattern is observed for pyridine-N-oxide( 2 9 ,4 0 ,4 1 ) and for triphenyl-
phosphines.(25 ,4 2 ) Pyridine' 4 3 -46) and imidazole(47 4- 8) experience o spin density
delocalization, although spin polarization introduces some spin density on the
7r system. When the aromatic ring is three bonds from the metal as in the case
of triphenylphosphineoxide,(49 ) or benzaldiminates,(50 ) or benzamides,(40 ) the
shifts still alternate but are about one order of magnitude smaller than in the
case of phenylamine type of ligands. In the case of salicylaldiminates, two pathways
sum up: one typical of the phenate moiety and one typical of benzaldiminates.
The latter gives rise to a smaller shift and lead to a system that is not alternating(3 9)
(Table 12-6). The porphyrins are another interesting ligand system that will be
treated later.
12-9 Factors Affecting Nuclear Relaxation in Paramagnetic Systems 519
12-9 FACTORS AFFECTING NUCLEAR RELAXATION IN

PARAMAGNETIC SYSTEMS
As seen in Chapter 7, nuclear relaxation is determined by fluctuating magnetic

fields, and the larger the coupling between the nuclear magnetic moment and
the fluctuating magnetic fields, the faster the relaxation. In general, we can write
T 1' = (E2/h 2 )f(re, w) (12-33)
where E2/h 2 is the square of the aforementioned coupling energy in frequency

units and f(re, w) is a function of the correlation time, r,, and of the Larmor
frequency, w. The correlation time, discussed in Chapter 7, is a time constant
typical of the system under investigation. Its reciprocal, 'E- , is the average rate
of change of the reciprocal position or orientation between the resonating nucleus
and any other particle whose movement relative to the resonating nucleus causes
the fluctuating magnetic field. The correlation function is of the form
ftre, w) = 1 2 (12-34)
1+ 0_2,
and provides the density of frequencies present in the system from the motion
described by the correlation time, c, The frequency w is that needed to induce
the nuclear transition.
The difference in a paramagnetic system compared to those in the previous
discussion of nuclear relaxation arises from the contributions from the fluctuating
magnetic fields caused by the unpaired electrons.
The unpaired electrons cause fluctuating magnetic fields by three main
mechanisms ( 3 ): (1) the electron relaxation itself, (2) the rotation of the frame
which contains the electron and the nuclear spins both aligned along the external
magnetic field (Fig. 12-13), and (3) chemical exchange involving bound and free
ligands (Fig. 12-14).
The first mechanism is self-explanatory. Electron relaxation occurs with T, 2
of the order of 10-5 to 10 - sec for radicals and 10- ' to 10 13 sec for metal
ions. From time to time, in an absolutely random manner, the electron spin
changes its orientation. In a large ensemble of electrons, the process of changing
orientation is a first-order process having a time constant T, 2 where T, and T2
are electron relaxation times. Here T1, 2 represents the correlation time for the
nucleus as far as this mechanism is concerned. The second mechanism leads to
a rotation of the electron magnetic moment directly around the resonating
B0
molecular frame holding the electron
and the nucleus together causes nuclear
relaxation. T, is the rotational correlation
I time.
B,
FIGURE 12-14 Binding and detachment

of the molecular frame, holding the
nucleus to the molecular frame, holding e-
the electron, causes relaxation. Tm is the
chemical exchange time.
nucleus. This causes a fluctuation of the magnetic field at the nucleus; the
correlation time is the rotational correlation time. The third mechanism is typically
produced by chemical exchange. A given molecule containing the resonating
nucleus binds to a molecule containing the unpaired electron and then dissociates
from it. The correlation time for this mechanism is the residence time in the
bound state.
As discussed in Section 12-2, we have spin density both on the resonating
nucleus and outside it that fluctuates with a time constant r, due to spin
reorientation. This is a fraction of an unpaired electron, equally distributed
between the Zeeman levels, which relaxes between the Zeeman levels just like a
full electron. Consideration of the existence of an excess population between the
Zeeman levels does not appreciably change the picture under these conditions
owing to the small extent of such an excess. (This is at variance with the isotropic
shift that arises only because of the excess Zeeman population.) When there also
is chemical exchange, spin density appears and disappears at the nucleus with a
time constant Tm. Thus, spin density causes fluctuating magnetic fields due to the
electron relaxation and in certain cases, to chemical exchange. The energy of the
coupling is the contact hyperfine coupling constant, A,/h, in frequency units.
Spin density outside the resonating nucleus, including the large spin density
at the metal ion where a large fraction of the unpaired electrons reside, is a source
of nuclear relaxation through dipolar coupling. The coupling energy now results
from an integral over all space of terms of the form of equation (12-22) summed
over the spin density distribution.(',"' 4) Upon rotation the integral averages to
zero; however, its square [see equation (12-33)] does not average to zero upon
rotation because (3 cos 2 y- 1)2 averages to 4/5. As with the isotropic shift, a
metal-centered contribution and a ligand-centered contribution (plus a cross
term) is considered.(2',s) The presence of spin density in a p, orbital of a carbon
atom affects the relaxation of a proton attached to the carbon.(5 6'5 7) The r 6
dependence of the average squared energy enhances the effect of the small fraction
of unpaired electrons on the nucleus when it is close to the resonating nucleus.
The ligand-centered contribution is often treated as a parameter.4 8 The relevant
nuclear relaxation mechanisms, arising from the dipolar coupling with spin
densities and unpaired electrons, are the electronic relaxation, the rotation of the
molecular frame, and chemical exchange.
When more than one effect is operative, the fastest one is dominant. Recall
the fastest process corresponds to the smallest correlation time. If the reorientation
due to rotation occurs for example with a time constant, (1/T), of 10~' sec and
that due to intrinsic electron relaxation occurs with a time constant of 10 - 13 sec,
the electron spin reorientation occurs on average one million times for a single
rotational reorientation and electron spin relaxation dominates.
We can define a correlation time, re- as
e =r, I +T 1 + 'm (12-35)
where r, is the rotational correlation time, r, the electronic correlation time, and
T. the residence time. In a more precise way, we should define z, and c,2 depending
on whether the correlation time is the longitudinal or transverse electron
relaxation time, but this distinction can be overlooked without serious errors.!2 ,s)
For different situations, each of the three mechanisms can dominate the dipolar
relaxation. In the case of certain copper(II) proteins r i > Z,- , Tr 107 sec 1
and TS-' dominates Z-1. In smaller molecules cr approaches 10" sec- and
dominates Te-1. The time-average coupling energy is zero in every case, whereas
its squared value is different from zero.
In the case of contact relaxation, molecular tumbling is ineffective for nuclear
relaxation and the correlation time is:
1
C
e = I + m (12-36)
We have a further source of relaxation that is relevant under certain

conditions. Let us return to the consideration of dipolar coupling in the absence
of chemical exchange. Whatever the correlation time, we have assumed that the
population of the Zeeman levels is equal. Correction for the Boltzmann population
difference does not change our conclusions regarding the mechanisms previously
discussed. However, this effect introduces a value of <S,> different from zero
giving rise to a permanent magnetic moment according to equation (12-13). This
small magnetic moment is seen by the nucleus as fluctuating with a correlation
time equal to r, If the correlation time for dipolar nuclear relaxation is also Tr,
the effect of the excess population is negligible. When the correlation time for
dipolar relaxation is dominated by Z 5 , r, is much longer, and the unpaired electron
(or a fraction of it) is reorienting with a time constant r,. On the other hand, the
permanent magnetic moment due to (S,> fluctuates, around the nucleus, with a
time constant r, (Fig. 12-15). The quantity <S,> is proportional to B, [Equation
(12-15)] and when r, is large, as with macromolecules in solution at high magnetic
fields where the excess Boltzmann population of the aligned state increases, the
latter rotation leads to an efficient relaxation mechanism. The relaxation
mechanism is also called Curie relaxation.

molecular frame holding the permanent
magnetic moment <Sz> and the
nucleus together causes nuclear
relaxation. Z, is the rotational correlation
time (cf. Fig. 12-13).
Equations for Contact Relaxation

Nuclear relaxation in the case of contact relaxation can be simply visualized in
the absence of chemical exchange. The transition between Zeeman levels of the
electron (i.e., the spin change) may cause nuclear reorientation. The transition
frequency required by the coupled electron-nucleus system is o0 - o0. This
frequency must be provided by the lattice. The equation for the longitudinal
relaxation(11) is:
2 A )2 2
T11 - S(S + 1) (12-37)
3 (h 1 + (or_- o, r
where the subscript M indicates the paramagnetic contribution, A, is proportional

to the interaction energy involving the spin density, equation (12-19), S(S + 1)
introduces the square of the intensity of the electron spin angular momentum,
o, is the nuclear Larmor frequency, a, is the electron Larmor frequency, T, is
2 2
given by equation (12-36), and (o1 - ot) can be approximated by o .
58
The transverse relaxation is given( ) by:
1 A (
T 2M-1 S(S+ 1) -2 c
+'(12 38)
2c2(
In contrast to Ti 1, T 2 1 contains a term differing from equation (12-34) in that

it contains o = 0 in the denominator because near zero frequencies affect T2
Equations for Dipolar Relaxation

For electrons localized on the metal ion and a resonating nucleus at a distance
r, the energy given in equation (12-22) averages to zero with time under any
2
modulating mechanism that is dipolar in nature. As mentioned before, E does
not average to zero. The correlation time r, is given by equation (12-35). The
equations for relaxation are:
2 9N2 N2 2/O B2
1M 15 (4)
-g-N2
h2r ( 1+
1)2 _2 2 1+
]7c
2
(12-39)
T - 1 - O 2 9N2 22:S(S+
2N 1) (1 c Icr
3 + 4E)
2M 15 (470 h 2r6 1 + o_ 2T,2 1 + o_12 2C
(12-40)
where gNPN h is p1 of equation (12-22) and gpBlh is P 2 . The frequency (o

corresponds to the nuclear Larmor frequency and o, is the electron Larmor
frequency. The latter is an approximation of o, - o (as in the contact term) and
of or + o,. The difference in frequency corresponds to the simultaneous flipping
of both spins, for instance, one decreasing by one unit in m, and the other
increasing by one unit in i. The transition is called a zero-quantum transition
and corresponds to W of Section 8-6. The sum in frequency corresponds to an
equal change in m, and i, It is a double-quantum transition and corresponds
to W2. As usual, T 2 contains a non-dispersive term without w and corresponds
to near zero frequency oscillations.
Similar equations can be derived for spin densities at specific atoms in the
molecule other than the metal ion.") These would be the equations for
ligand-centered relaxation, which would require knowledge of the spin density
distribution in the molecule and the introduction of cross terms with the
metal-centered relaxation equations.
Equations for Curie Relaxation

We have already seen that a paramagnetic molecule has a magnetic moment
given by equation (12-13). Upon rotation, this magnetic moment causes a
fluctuating magnetic field at the resonating nucleus and causes relaxation (Fig.
12-15). This mechanism differs from those above because it depends on (S,>. Its
effect depends on the square of the magnetic moment(5 9)and according to equation
(12-15) on the square of the applied magnetic field.(60 -61 ) The correlation times are
Ze- = ,M + r' (12-41)
for the dipolar coupling and
'e = (12-42)
for contact coupling. These equations differ from equations (12-36) and (12-37)
in that the term r, is absent. Otherwise the same considerations hold. For usual
fields and small molecules, the effect is negligible. The rotational correlation time
can be predicted for a spherical molecule through the Stokes-Einstein equation(2):
3
4r, =
(12-43)
3kT
where q is the viscosity of the solvent and a the radius of the molecule. Consistent
with this formula, r, for a hexaaqua complex is about 3 x 10- " sec and that
for a protein of molecular weight 30,000 in solution about 10 -8 sec. The expression
for T2 ' for the dipolar coupling between the permanent magnetic moment and
the resonating nucleus (dipolar Curie relaxation) is( 60 -61 :
1 o 2 (0 2 94 YB4s2(S + 1)2 3T,

T_' 2- 5 (47n) (3k T)2r 4T, + 1 + W2,2 (12-44)
where o is gNpNBo/h. Since the line width is proportional to T

2
T = 1 Av('/ 2 ) (12-45)
where Av('/2) is the line width at half height, the Curie relaxation can affect the
line width. At 600 MHz, the protons within 10 A of a spin S = 3/2 metal ion and
with a Zr of 10 - sec are broadened. They are broadened beyond detection when
r is smaller than 5 A.
The expression for T 2 '[equation (12- 14)] contains in the f(re, w) part, a
term in re. Relaxation mechanisms whose correlation time is long largely affect
T 2 '. For T ', there are only terms of the type T/(1 + o 2z2 ) and for large T values
they tend to zero.("') In both cases, if T, is small, the effect is small.
General Comments Concerning Relaxation

All the foregoing equations have been derived with the assumption that the S
manifold is not split at zero magnetic field but is only split by the magnetic field
into the 2S + 1 levels of the S multiplet. However, we know from epr spectroscopy
and from magnetic susceptibility measurements that this is rarely the case. Zero
field splitting is usually present when S > '/2. Furthermore, hyperfine coupling
causes splitting of the S manifold at zero magnetic field (see, e.g., Fig. 12-2) and
the splitting is sizable when the coupling occurs with magnetic nuclei which bear
the unpaired electrons. These splittings do not have an effect on nuclear relaxation
as long as the splitting energy E,, is smaller13 1 than he -. For example, for
e- 10-" sec, E,, has to be smaller than 1 cm-1 When this is not the case,
the foregoing equations provide qualitative predictions, and when quantitative
5 62
treatments are needed the reader is referred to available literature.(2, 3, -71)
As in the case of the isotropic shift, the factorization among the various
contributions is the first step in any quantitative exploitation of the relaxation
parameters. Although the nature of the problem is complicated, we are fortunate
in this case because the various contributions may differ by orders of magnitude.
Therefore, in general, only one contribution is dominant and simple consideration
and sample calculations can show which mechanism dominates. For example,
T1 of nondirectly coordinated hydrogen atoms in paramagnetic metal complexes
is determined by dipolar contributions, which sometimes may be approximated
by a metal-centered model. Nuclei of atoms directly coordinated to paramagnetic
centers may have large contact contributions to T, and T2 . In the case of the
proton nmr of copper complexes, T, is determined by the dipolar contribution,
whereas T2 , which contains the non-dispersive r, term, is determined by the
contact contribution.17 2 ) In fact, the correlation time for contact relaxation is zT,
which is ~10-' sec, whereas the correlation time for T, is r,, which can be as
short as 10-" sec.
12-10 RELAXOMETRY
The equations for nuclear relaxation by unpaired electrons always contain either
or both the nuclear and electronic Larmor frequencies. Such terms make the
nuclear relaxation magnetic field dependent. A single relaxation measurement
does not provide both the correlation time and the electron-nuclear coupling
energy. In the ideal situation, enough experimental data would be available at
different magnetic fields to describe the curve for the entire field dependence of
T1 or T2A
The expected plots of Til and T 2 ' as a function of the applied magnetic
field are reported in Fig. 12-16 for the contact and in Fig. 12-17 for the dipolar
relaxation.(2.53,5 8 5, 9 Experimentally, at the time of writing, we are limited on the
high field side to 600 MHz, whereas on the low field side the limit is the sensitivity
which is proportional to the strength of the applied magnetic field, to the nuclear
magnetic moment, and to the concentration of nuclei. Devices exist for measuring
12-10 Relaxometry 525
2-rc
FIGURE 12-16 Plot of (A)

B the f,(w, r.) function in
I.. Tc equation (12-37) and (B) of
the f2(o, Tc) function of
equation (12-38) as a
function of the magnetic field.
0.5re -
103 10, 105 108 1 108 109 1010

Proton Larmor frequency (Hz)
20Te
) 15te -
FIGURE 12-17 Plot of (A)
the f,(o, ze) in equation
(12-39) and (B) the f2 (0, Tc)
function of equation (12-40)
as a function of the magnetic
field.
S5tc- A
0
103 10' 105 106 101 108 109 1010
Proton Larmor frequency (Hz)
Ti, and sometimes T2 , of solvent nuclei, particularly water hydrogens, at very

low magnetic fields.(', 3,73,74)
?
The problem of chemical exchange as a source of nuclear relaxation has

been neglected in this chapter except when ',, becomes the correlation time. When
T. ' is larger than the separation in Hz (sec-1) between the signal in the A site
and that in the B site, then only one signal is observed whose shift is the average,
weighted for the occupancy of the two sites. The treatment is analogous to that
in diamagnetic systems. If m1 is larger than the difference in both Ti 1 and T 2
then the resulting relaxation values are also a weighted average. When Tm ' is of
the order of or smaller than the relaxation rates and the chemical shift difference,
then more specialized literature should be consulted. 2 ,7 s-78) Here we assume
that chemical exchange is fast and is averaged by the mole fraction relation.
We will treat coordination compounds in aqueous solution having a water
molecule in the coordination sphere. The water proton T, values are measured
in a range of magnetic fields between 0.01 and 100 MHz. First, we shall consider
FIGURE 12-18 T2 1 (o) and Zuu

T (o) of water protons in
6.65 x 10' M Cu2 * solutions
as a function of the proton 100
Larmor frequency at 25 C.
The continuous lines
represent a fitting of the data
according to the Solomon 0
equations (12-40) and
79
(12-39), respectively.()
Proton Larmor frequency (MHz)
the complex Cu(OH2) . The experimental data reported(79 ) in Fig. 12-18 are
a part of the curve reported in Fig. 12-17(A) for dipolar relaxation. By fitting
the curve with equations (12-39) and (12-40), a value of z, of 3.0 x 10-" sec is
obtained at 25 C. This is a typical time constant for rotation of a hexaquacomplex.
Since the correlation time is the rotational time, the nuclear relaxation must
depend on dipolar coupling for only this effect depends on rotation. Assuming
that the measured effect is entirely due to the presence of water in the first
coordination sphere, and that there are 12 equivalent protons, the metal-proton
distance is estimated to be 2.8 A, in agreement with the expected 2.1 A
metal-oxygen distance.
These conclusions show that the proton-metal coupling is a metal-centered
coupling for which equation (12-39) holds. The validity of the metal-centered
point dipole approximation has been shown to be purely accidental as a result
of cancellation of terms due to the delocalization of the electron within the Cu-O 6
volume and that due to spin density on the oxygen atom.(54 ) Proton contact
253
contributions in all the water-copper(II) interactions are negligible.( )
When the solution is made more viscous by adding ethyleneglycol or when

the copper ion is bound to a macromolecule, Tr, increases. For a protein of
molecular weight 32,000, such as copper-zinc superoxide dismutase (SOD
hereafter), 7, is expected to be around 3 x 10-8 sec. The rotational mechanism
does not determine -c any more because r, is now longer than either or both -,
or .. Furthermore, there is a splitting of the S manifold at zero magnetic field
due to the coupling of the electron with the magnetic nucleus of copper. Such
splitting is of the order of 10-2 cm-' and is smaller than hre- when re is
3 x 10 sec (equal to 0.2 cm- '). The splitting is larger than he ' when -e is
of the order of 10-' sec. The experimental data for an aqueous solution of SOD
are reported(s01 in Fig. 12-19. Their fitting to the dipolar coupling model,
including the coupling between the electron and the copper nucleus, has provided
a re of 3 x 10-' sec a copper oxygen distance of 2.4Afor a single water molecule
in the coordination sphere.1 65 ) The value of Te could correspond to the water
exchange rate, rT, or to the electronic relaxation time, r,. It is reported that r,
corresponds to r,.1 6 5 Relaxometry is, therefore, a tool to determine the electronic
relaxation times, provided that rotation is slower than electronic relaxation.
4
25 FIGURE 12-19 Water proton
E3
2 relaxation data for superoxide
dismutase solutions.80 The solid
line is the best-fit curve
obtained with the inclusion of
the effect of hyperfine coupling
CL - with the metal nucleus.6 5 ).
0.01 0.1 1 10
Proton Larmor frequency (MHz)
12-11 ELECTRONIC RELAXATION TIMES
The electronic relaxation times estimated with relaxometry are reported in Table
12-7.25"' They are referred to a magnetic field of 2.35 T, (100 MHz 'H resonance
frequency) because electronic relaxation has been also found to be field dependent
in some cases. These relaxation times are very relevant for understanding the
nmr of paramagnetic molecules. In the cases of dipolar relaxation dominating
TABLE 12-7. Electronic Relaxation Rates for Various

Metal Ions and Nuclear Line Broadening Due to Dipolar
Relaxation"
Nuclear Line
Metal Ion ' (sec 1) Broadening (Hz)"
Ti3 + 109 -101 3000-500

VO 2 + 108 l 109 20,000-3000
V3+ 2 x 10" 100
V2+ 2 x 109 9000
Cr3+ 2 x 109 9000
Cr210'' 300
Mn3+ 1010 - 10" 3000-300
Mn2+ 108 - 109 200,000-40,000
Fe 3+ (H.S.) 1010 - 10" 5000-400
Fe 3+ (L.S.) 10" 102 40-10
Fe" (H.S.) 102 70
Co 2 + (H.S.) 10" - 10 2 200-50
Co 2 + (L.S.) 109 - 1010 3000-500
2
Ni + 1010- 102 1000-25
Cu 2 + 3 x 101 - 109 9000-3000
Ru 3 + 10"- 102 10-40
Re310'' 100
Gd3+ 108 - 109 400,000-60,000
Ln3+ 1012 30-100
"For 'H, r = 500 ppm, B0 = 2.35 T.

and r, being the correlation time, the line width of a proton at a distance of 5 A
from the metal is predicted from equation (12-40) and reported in Table 12-7.
From inspection of the table, it is seen that some metal ions, including most
of the lanthanides, give rise to little nmr line broadening. Low spin iron(III), high
spin octahedral cobalt(II), and high spin tetrahedral nickel(II) fall in this category.
Manganese(II), gadolinium(III), copper(II), and vanadium(IV) give rise to very
broad signals. The first class of metal ions are called shift reagents because they
provide hyperfine shifts with little broadening, whereas the latter class is called
relaxation reagents because they relax the nuclei. In excess ligand and under fast
exchange conditions, the line broadening, or Ti ' enhancement, is reduced
according to the mole fraction of bound ligand. Under these conditions, a ligand
nucleus can be observed in the presence of a relaxing agent yielding fruitful
information from the relaxation parameters.
Knowing the values of T, is a starting point to attempt an understanding of
the factors that determine electronic relaxation in solution. Every time r, is shorter
than rT,we can propose that fluctuating magnetic fields associated with rotation
cause electronic relaxation. In addition to the intensity of the magnetic moments,
the electron spin differs from the nuclear spin because the electron spin angular
momentum is coupled with the orbital angular momentum. The angular
momentum is orientation dependent. This leads to g values and zero field splitting
values that are orientation dependent. We say that rotation modulates the g
anisotropy and the zero field splitting. Such modulations are associated with
modulations of the orbital magnetic moment and cause electron spin relaxation.
When rotation is slower than electron relaxation, rotation cannot be the
mechanism that leads to electron relaxation. In this case, it is instructive to refer
to solid-state mechanisms. Indeed, lanthanides have comparable electron relax-
ation rates in solution and in the solid state, which are much larger than rotation
rates. It may be assumed that on the time scale of electron relaxation the
surrounding solvent of a metal ion in solution can be visualized as a small crystal.
In this context, we shall discuss some results in solids.
In solids, there are crystal vibrations of low energy whose excited levels are
populated according to the temperature.""1 ) Changes in the excited vibrational
levels cause modulation of the metal-donor distances, which are reflected in
changes of the orbital angular momentum, which in turn lead to electron
relaxation. Differences in crystal vibration frequencies due to population or
depopulation of a given level are frequencies called phonons.(2' In crystals there
are continuous emissions or absorptions of phonons through exchange with other
properties of the solid (e.g., with the electron spin levels of a paramagnetic
substance). Around room temperature there are many phonons with frequencies
of the order of 100-1000 cm-'. Therefore, if the substance has electronic spin
levels with energy separations of this order of magnitude, electron relaxation will
be very efficient.(8 3 ) This is the case of lanthanides where the ground levels reported
in Table 12-1 are split by the above order of magnitude from the action of the
crystalline field.(s4 ) The same situation holds for high spin octahedral cobalt(II)
and tetrahedral nickel(II) complexes.!") The ground state is a "Tg and a 3 T,
respectively (i.e., it is orbitally triply degenerate). Low symmetry components and
spin orbit coupling give rise to six Kramers' doublets in the former case and up
to nine levels in the latter. Among these, energy separations of the order of the
phonon energies exist at room temperature. Octahedral nickel(II) and tetrahedral
cobalt(II) have an orbitally non-degenerate ground state so the low symmetry
component and spin orbit coupling provide smaller values of the zero field Lattice Spin Lattice Spin
splitting. It follows that electron relaxation is slower because there are fewer
phonons of that energy.
The general idea is that, whenever there are low-lying excited states, phonons fT
can provide the energy for transitions involving a change in m, (Fig. 12-20(A)).
This mechanism is called the Orbach process. When there are no excited states
at energies around 100 cm -', the transitions between Zeeman levels occur through (A)
a combination of two phonons whose energy difference matches the energy - -b b

required for electron relaxation. This process, called the Raman process1 84 ' 86) (
[(Fig. 12-20(B)], is far less efficient and accounts for the long electron relaxation (B)
times of copper(II), vanadium(IV), high spin manganese(II), etc. In the case of
FIGURE 12-20 (A) Orbach
high spin iron(III), (second order) spin orbit coupling is rather efficient in splitting
the 6 S. When there are strong anisotropies as in the case of some porphyrin and (B) Raman processes. In
an Orbach process, the
complexes, the splitting is of the order of several tens of wavenumbers and the
electron spin is excited to
'H nmr signal can be observed. In general, the electron relaxation rates of high level c by absorption of a
spin iron(III) complexes are difficult to predict because zero field splitting is lattice phonon and then
variable. emits another phonon to
In summary, electrons can relax through the rotational mechanisms for level b. In a Raman process,
componds with slow electron relaxation rates in solution. When rotation cannot the electron spin is excited
account for the observed values of electron relaxation, we should refer to the directly to an excited level by
simultaneous absorption and
emission (scattering) of a
single phonon.
2
(C) H NMR
la-d 2 in H2 0
H H
N/ H 2
oN 5 25Hz
cu4
CH,
FIGURE 12-21 (A) 'H nmr

40 30 20 10 0 spectra at 23 C of the
ppm compound shown and (B) its
partially deuterated form in
(B) H NMR m
D2 0 (shaded arrow denoted
the resonance position of the
la-d
32 in D2 0 5,
4
CH 3
2-H signal) and (C) 2 H nmr
spectrum at 23 C in H20.
(A)' H NMR
lain D2 0
I I I I II 1
30 20 10 0
type of mechanisms observed in the solid state. The presence of unpaired electrons
always causes line broadening. Equations (12-38) and (12-40) provide an estimate
of the line broadening as a function of r,; the latter may or may not be equal to
the electronic relaxation times. In any case, the electronic relaxation times
represent an upper limit for the broadening. For example, a proton at 5 A from
copper(II) would experience a broadening of ~5000 Hz at 100 MHz (Table 12-7),
if the correlation time were determined by T, This holds for macromolecules that
rotate slowly. In small complexes, however, T, is determined by a short r, and
therefore the line broadening is smaller. Finally, a proper choice of the magnetic
field should be made by taking into account equation (12-44) (Curie relaxation).
A comment is also appropriate regarding the choice of the resonating nucleus.
The metal nuclei bearing the unpaired electrons cannot be observed because the
coupling is too strong and the line width too large. Nuclei with small magnetic
moments are broadened less because the coupling energy is small (although the
magnitude of the hyperfine shift decreases when expressed in Hz). It is common
to use a deuteron instead of a proton when the linewidths of the latter are too
large to be detected. In Fig. 12-21, the 'H and 2 H spectra of copper(II) complexes
are shown.(8 7 )
12-12 CONTRAST AGENTS
NMR imaging is a new clinical technique to obtain an image that is used

particularly for the head and spine. The technique is sensitive to the protons of
tissue water and fat in both soft tissue and bone marrow, and is insensitive to
bone itself. Proton density varies only slightly among tissues. The change is not
large enough for images of proton density alone to be clinically useful. However,
the relaxation rates of tissue protons vary significantly from one tissue to another,
so that contrast in the nmr image depends mainly on the relaxation rates, T,
and T 2. The contrast can be optimized case by case with the appropriate weighting
of the relaxation rates.(88)
It is the aim of researchers in the field to improve contrast and resolution
for a given image in order to obtain a better image or a comparable image in
less time. A reasonable way involves altering relaxation rates by introducing a
paramagnetic compound into the tissue, which acts as contrast-enhancing agent.
From the inorganic point of view, a contrast agent is a complex with a
capability of dramatically increasing the proton Ti-' and T 2 '. According to
equations (12-37) to (12-40), the complex should have a large S and a long z,.
These properties are met in Gd 3 1 (S = 7/2), Mn 2 ' and Fe3 (S = '/2). Further-
more, water molecules should be in fast chemical exchange conditions with the
paramagnetic center as is usually the case for water molecules coordinated to
Mn 2 + and Gd 3 . The correct approach requires measurement of T, and T2 of
a water solution of the complex at various magnetic fields in order to check the
relaxivity (solvent relaxation rate enhancement per mM solute) at the imager
frequency.( 89) Then the complex should be tested in tissues and the water proton
relaxation rates remeasured directly at different times after the injection. Tissues
behave essentially as liquids.(88) Experiments in tissues can be made in vivo in
test animals or after surgical asportation of pathological samples.
The organic part of the ligand is responsible for targeting a certain tissue
when the contrast agent is administered as a drug. At present, the contrast agent
CH2COO(-)
OOC - H2C CH 2 - COO- FIGURE 12-22 The DTPA5 -
N- (CH2)2 - N () anion.
OC H2 C-N - (CH2 )2
gadolinium (diethylenetriamine pentaacetate) 2 - (GdDTPA2-, with DTPA5

shown in Fig. 12-22) is commercially available. Other gadolinium complexes are
being investigated.'90 )
12-13 TRENDS IN THE DEVELOPMENT OF PARAMAGNETIC

NMR
One main goal in nmr is the assignment of peaks. Ti ' and T 2 ' are often
dominated by dipolar coupling and, therefore, depend at least qualitatively on
the sixth power of the nucleus-metal distance. These measurements, together
with the analysis of the shifts and some experience, can lead to an assignment.
The relative intensities of the signals and substitution of a nucleus, often a proton,
with another group or isotope help in the assignment. NMR in diamagnetic
systems has gone far in establishing criteria for assignment and in relating nuclear
properties with internuclear distances. The trends in nmr of paramagnetic
molecules parallel the applications of the well-established techniques in diamag-
netic systems, with the limitations imposed on the paramagnetic system from fast
nuclear relaxation.
The Nuclear Overhauser Effect

The nuclear Overhauser effect (nOe), in its more classical form, is the fractional
variation of a signal intensity when another signal is saturated for a long
time.?1' 9 2 1 This is called steady-state nOe (see the diamagnetic nmr chapter,
Section 8-6). If the nuclei corresponding to two signals interact through dipolar
coupling, then a steady-state nOe is observed which, as already discussed, is given
by the equation
- (12-46)
2W 1 + W2 + Wo p
where a is the dipolar cross relaxation and p can be taken as T 1 '. Here, a
corresponds to the net magnetization transfer from the saturated nucleus to its
neighbor under observation and is given by
2
pu 61
r 1+ 4.o
2
,,
2 (12-47)
where p and U2 are the nuclear magnetic moments, r is their distance, and T, is
the correlation time for their reciprocal reorientation.
Paramagnetic molecules are characterized by large p values and therefore
j is small. For protons, in the ortho position of aromatic rings or geminal in a
CH 2 moiety, the distance is small and even if p is of the order of 102 sec- ,
can be of the order of 1 to 50%. Therefore, nOe can be measured under high
signal-to-noise conditions. The conditions for observing nOe are best if the
magnetic field is high and r, is large. Therefore, applications are successfully made
in the study of paramagnetic metalloproteins where u, is large corresponding to
the rotational correlation time, equation (12-33). Some examples of nOe
experiments in different systems are shown in Fig. 12-23.( 3'94)
Steady-state nOe in diamagnetic molecules result in the observation of
secondary nOe through spin diffusion. The effect is greatly reduced in paramag-
netic systems where nuclear relaxation is so fast that secondary nOe do not have
time to build up.(9 5)
Transient nOe is the fractional variation of a signal intensity when a related
signal is inverted with a 180 pulse. Since the inverted signal returns to equilibrium
with a time constant T ', which is large in a paramagnetic system, magnetization
transfer is small. In general, transient nOe experiments give poorer results than
steady-state nOe experiments.(95 ) Nuclear Overhauser effect measurements pro-
vide information on internuclear distances and, therefore, provide a unique
criterion to perform a scientifically correct assignment.
The Effect of Fast Nuclear Relaxation on 2-D Spectra

The 2-D experiments, COSY and NOESY (see Section 8-19), are also possible
for paramagnetic systems. In this case, the limit is that the magnetization transfer
fg h k np
1 ++ ox
FIGURE 12-23 Some h

examples of nOe difference
spectra of paramagnetic h ox
metalloproteins. (A) Reduced 3 +
ferredoxin from P. umbilicalis
in D2O(3) (B) N3 adduct of
copper(Il)-cobalt(II)
superoxide dismutase in
H20. The nOe's are typically fg h
of a few percent.
CF
4
5x5
9 *
I II 11 1 1 I I 1 I I II I I I 11 11
IIIII111 Ii
35 30 25 20 15 10 5 0 -5
ppm
(A)
or redistribution responsible for cross peaks should occur before the spin system
returns to the starting equilibrium situation. The return to equilibrium is governed
by the relaxation times T, and T2, which, as we have said, may be very short.
Sometimes, T and T2 are so short that there is not enough time to alter the
magnetization of the coupled spins. Recent experiments have shown that for T
or T2 of the order of 50 msec, 2-D experiments are feasible with millimolar
solutions. The results are useful if other conditions, like a long r, for NOESY
96
experiments or large J values for COSY experiments, are satisfied.( 99) One
example is reported in the following section.
G
KL
M HN R
C
F0
2 . .
K
A
FIGURE 12-23 (continued)
LP R
120 90 60 30 0 -30
8(ppm)
(B)
R
12-14 SOME APPLICATIONS
Planar-Tetrahedral Equilibria in Nickel

X Ni/2 One of the early applications of 'H nmr to paramagnetic complexes involved his
N N-substituted aminotroponeiminate nickel(II) complexes.. I A chemical equi-
librium exists between planar, diamagnetic, and tetrahedral, paramagnetic
R
species."') Peak assignments were performed on the basis of chemical substi-
tution and on the analysis of the first-order J splitting that is observed when the
diamagnetic species is dominant. The shifts on the ring are alternating, as expected
for a 7E delocalization pattern."o 2 ) A large variety of substituents, R, on nitrogen
were investigated. When R is aliphatic, the shifts decrease with the distance from
the paramagnetic center, whereas in the case of aromatic substituents large shifts
Ni Ni are also observed for protons very far away (Table 12-8).("')
Since the planar-tetrahedral equilibrium is fast on the nmr time scale, the
isotropic shift depends on the molar fraction of the paramagnetic species. The
equilibrium constant is expressed as
K =[] 0
[D] 1 -
TABLE 12-8. 'H Contact Shifts or R Groups in N-

substituted Nickel Aminotroponeiminatesa(1 0 3)
R 6169
CH 3 3.36
CH 2-CH 3 CH 2 3.24
CH 3 0.21
CH2-CH2 CH 3 CH 2 3.44
CH 2 0.23
CH 3 0.12
0o-0.36
/ \0.38
p -0.51
0o-0.37
H, m 0.37
CH 3 0.63
0o-0.37
m 0.38
/\1 -\o0.037
-m 0.051
p -0.059
0 -0.39
rn 0.34
o1 0.01
- - m -0
p -0.01
The shifts are relative to the shift of the # proton of the troponeiminate
ring, o.
6
Such shifts are in the 20 to 30 ppm range.
where [P] is the concentration of the tetrahedral paramagnetic species and [D]
is the concentration of the planar diamagnetic species.( 36 > The molar fraction of
paramagnetic species is given by ot:
(AV)observed
(Av)fuIIy paramagnetic
When Av of the fully paramagnetic species is not available, one should rely on
10 4
the validity of equation (12-17) for estimating such a value. (
Diastereoisomerism
Bis-N-substituted salicylaldiminates also undergo planar :± tetrahedral equilibria.
1 05
Studies similar to those of the aminotroponeiminates have been performed. )
The bis-chelate complexes are dissymmetric and can be A or A. As the
R = sec-butyl group bears a center of asymmetry, the bis N-sec-butylsalicyladimi-
nate complexes can exist as two diastereoisomers A + and A -, which are pairwise
equivalent to A- and A +, respectively.( 1061 ) The two diastereoisomers are
detected thanks to sharp lines and to the two different isotropic shifts (Fig. 12-24).
TMS
act.
H Ni/2
X= 0 H CH CH2 CH,
-H
1
CH CH3 FIGURE 12-24 100 MHz H
meso meso nmr spectra of mixtures of
act. act. active and meso isomers of
Ni(sBuPhHH) 2(X = 0) and
0 IMS Ni(sBu-SPhHH) 2 (X= S) in
11 Ho TMS CDC13 solution at -300.
+ + Chemical shifts are given in
Hz relative to TMS. 106 )
X S a-H CH p-H
meso meso meso
act. act. act.
| I + +
Diastereoisomerism and Diastereotopism in Cobalt(i)

N-alkyl pyridin-2-aldimines give rise to tris-chelate cobalt(II) complexes. Due to
the asymmetric nature of the ligand, cis and trans isomers are expected and
observed (Fig. 12-25).("') The contact and pseudocontact contributions to the
shifts lead to a large difference in shifts for the various isomers. When R is
CH3
CH3
R =-CH
Hy
CH3
1
FIGURE 12-25 H nmr
spectra at 26 of
Co(N-isopropyl
pyridin-2-aldimine) 32 in cisCH, cisCH 3
d3-acetonitrile (top) and
d6-acetone (bottom).(1 07
FU X
I I | |
20 10 0 -10 -20 -30
S(ppm)
CH
3 isopropyl, the two CH 3 groups are inequivalent because they sense an asymmetric
I center that is the metal. Two signals are expected for the cis isomer and six for
the trans isomer. Indeed, eight well-separated signals are observed. In every case
the intensity of the cis signal is minor corresponding to a less stable isomer (Fig.
12-25).(107)
0J 0NGroup Inequivalence in NiSALMeDPT

The five coordinated, bis (3,3'-salicylideneiminatopropyl) methyl amino nickel(II)
complex is asymmetric, with the CH 3 group making the two salicylaldiminate
moieties inequivalent when coordinated. The signals of the two
salicylideneiminatopropyl groups will differ (as seen in Fig. 12-26). This system
was first studied by LaMar and Sacconi,' 08) who related the difference in shift
CH,
3 N N __
a 4 H
-56 Ni ,X
CH -N
Ni-N CH
0 to the existence of a more square pyramidal (almost equivalent groups) or more

N trigonal bypyramidal (more inequivalent groups) isomers. The generality of the
inequivalence in one isomer was then shown for the bis thio analog." 0 9) The
assignment of the aromatic signals was performed by varying substituents. Next,
the spectral width was increased to include all the signals.') Finally, T,
measurements and 2-D spectra have led to a complete assignment.!96 )
6H
Fl
4H
6H
3H N-CH3
p-CH n 4H
a-CH L- CH
I I | 3-H
30 20 10 0 -10
ppm
y-CH
CH =N a-CH aC
I | I | I |
500 400 300 200 100 0
ppm
1
FIGURE 12-26 Room temperature H nmr spectra of 5C-NiSALMeDPT in CDCl3-
Each proton in the molecule (except the methyl protons) gives rise to a separate signal.
Ion Pairing
Ion pairs involving a paramagnetic anion and a diamagnetic cation have been
investigated through 'H nmr." 0 "") The protons of the tetraalkylammonium
cation experience an isotropic shift, which is probably contact in nature
(Fig. 12-27). When the pair is formed, it appears that spin density is transferred
onto the diamagnetic cation. The approach of the two anions occurs from every
possible direction, and the free ± paired equilibrium is fast on the nmr time scale.
Therefore, all the equivalent cation protons remain equivalent upon pairing.
These systems were investigated through the shifts(i 101 11) and T, values." 1i) The
formation of ion pairs is enough to provide orbital overlap and transfer of spin
density. Cases can be conceived where a moiety can approach a paramagnetic
system at van der Waals distance (i.e., the spin transfer is negligible, but the
nucleus still senses the unpaired electron through space). In this case, the
correlation time for the interaction can be the diffusion correlation time 'I and
the relaxation equations have a form of the type(112,1 13)
Ti = (32r/405)?r2ys2h2S(S + 1)(NA/1000)([M]/(d/D)) - (J 2(OS - OI)

+ 3J1 (w1 ) + 6J2 (ws + or)] (12-48)
T2 (167r/405)yi 27s2h2 S(S + 1)(NA/1000)([M]/(dD)) - [4J,(O)

+ J 2(ws -- o) + 3J(r)+ 6J 2 (ts + co) + 6J 2 (ws)] (12-49)
where NA is Avogadro's constant, d is the distance of closest approach, D is the

sum of the diffusion coefficients for the two molecules, and [M] is the
FIGURE 12-27 A contact

interaction occurs in MX4
(NR4)+(MX 4)- ion pairs involving
the a-CH 2 groups of the
cation 1 10,111)
H HH H H H
\/ \/ \/
C C C
(CH2)2 (CH2)2 N® (CH2)2

II I I
OH3 CH3 (OH2)3 CH,
CH3
concentration of the molecule bearing the paramagnetic center. Under a wide

range of conditions the spectral density functions are given by
4
J = J2 = (1 + 5z/8 + z2/8)/(1 + z + Z2/2 + z'/6 + 4z /81 + Z1/81 + Z6/648)
(12-50)
with
z = (2wD) 1/2 (12-51)
In turn, TD is given by
2d 2
TD = (12 52)
D
A model of this type has been used to describe the interaction of water with
3
Cr(en)3 + (114)
Hemin-Imidazole-Cyanide
The spectrum of a heme complex (i.e., hemin-imidazole-cyanide) is an example
of a low spin (S = 1/2) iron(III) heme system. Heme complexes are present in
many proteins and studies have led to understanding of these important complex
systems. The spectrum of a model is shown in Fig. 12-28).(".5)The four intense
signals are assigned to the four ring methyls. Their precise assignment can be
Hr' o
" 'P
3- Ha
H3 C 12 - / 34 C HP
N N
H 4 e H
H3 C 8 y 6e CH 3
aCH2 H aCH 2 FIGURE 12-28 Room

I I temperature 600 MHz 1Hnmr
PCH 2 CH 2 spectrum of
3-CHI 1-CH 3 I I hemin-imidazole-cyanide.
COU2 co 2 The hemin scheme is
8-CH 3 4,5 reported in the inset.
2'
5-CH 3 6-Ha meso a
6-Ha' meso p
7-Ha meso 8 2-Hp3
7-Ha' 6, 7-Hp 4-Hpt
2-Ha meso y 6, 7-H1p 2-Hp.
14-H 14-HP
5-114 4-H-H p
i I I I
16 14 12 10 8 6 4 2 0 -2
made by using selectively deuterated samples. Alternatively, 2-D NOESY and

COSY spectra can be employed. In the NOESY spectra, cross peaks occur between
adjacent protons. In the COSY spectra, the cross peaks occur between signals
related by J coupling. Vinyl and propionyl group correlations are found both in
NOESY and COSY spectra, as shown in Fig. 12-29!(..5) Methyl groups are
distinguished because 1 and 3 give a nOe with vinyl groups. Finally, methyls I
and 8 give a nOe to one another. The methyl signals, once assigned, give a nOe
with the meso protons and, therefore, the spectrum is fully assigned. As far as 50N
the imidazole is concerned, the 1'-H signal disappears when the spectrum is- -
recorded in D 2 0 and 2'-H as well as 4'-H have the shortest TT All these spectra i
were recorded at 40 mM hemne concentration." 15) in the case of proteins, the
concentration is about one order of magnitude lower and, as a result, some cross
peaks may be lost. c
Spin Delocalization in Iron Porphyrins

Iron porphyrins are important molecules especially for their biological relevance.
Iron(III) can be high spin or low spin, although in some cases spin admixtures
are proposed. The electronic levels in the two cases are shown in the following
oj ~ti
hs
g -_
or
FIGURE 12-29 nq
Two-dimensional 1H COSY
(upper left) and NOESY
(lower right) spectra of
hemin-imidozole-cyanide.
The labeling on the spectrum
(lowercase = COSY, p
b
uppercase NOESY) *d m dNOOR
indicates the following
correspondences to the
scheme in Fig. 12-28.
Yc5~
e, y Y2
. CDO Ve
oAB OU
9
ol| I | | | |
16 14 12 10 8 6 4 2 0 -2
ppm
8-CH 3/7-Ha 2-Hc/2-Hfl, P: 7-Ha/7-Ha' 3-CH 3/meso a

8-CH 3/7-Ha' 4-Ha/4-H#, q/Q: 7-Hoi/7-H#,l' 1-CH 3/meso 3
5-CH 3/6-Hac 4-Ho/4-Hfl, r/R: 7-Hoc'/7-H#l,#l' 2-Hc/meso La
5-CH 3/6-Ha' 2/4,5 s: 2-Hflt/2-Hfl, 4-Hc/meso #
3-CH 3/4-Hot 6-Ha/6-Ha' t: 4-Hf#1/4-Hfe 3-CH3/4-H#,
1-CH 3/2-Ha 6-Ha/6-Hfl,# U: 8-CH 3/meso 6 1-Ch,/2-H#,
2-H /2-H#, 6-Ha'/6-H# V: 5-CH 3/meso f#
scheme. High spin complexes have unpaired electrons in --type orbitals." 1 6 )
Therefore, a dominant a delocalization and large downfield shifts are expected.
The pyrrole protons of meso-tetraphenylporphyrin (TPP) chloride complexes are
far downfield (Fig. 12-30).(2,116)
+
d,, d,
+
+ d,
D4 D (elongated)
High spin Low spin
Similar considerations hold for high spin iron(II). In contrast, the unpaired
electron is in a it-type orbital in the low spin iron(III) TPP bis-imidazole
complex.f(16 ) The shifts are smaller and the pyrrole proton is shifted upfield (Fig.
12-30). Since low spin iron(III) complexes have short electronic relaxation times,
p-H
o-H FeTPPCI (S = 5/2)
Pyrrole H
rnm-H
FIGURE 12-30 Simulated 1H

nmr spectra of
1-H 5-H meso-tetraphenylporphyrin
o, m, p-H (TPP) iron complexes.(')
Chemical shifts are in ppm
Pyrrole H FeTPP2-CH 3 -lm (S= 2) from TMS.
2-CH3
m, p-H
o-H
Pyrrole H
FeTPP (S= 1)
I I i I I I I I I 1
90 80 70 60 50 40 30 20 10 0 -10 -20
8(ppm)
the nmr signals are quite sharp. The 'H nmr spectra of iron(III) hemin-
imidazole-cyanide,(1 15) shown in Fig. 12-28, provides a typical example of this
class of compounds.
Cobalt-substitutedCarbonic Anhydrase
Carbonic anhydrase (CA) is a zinc protein of MW 30,000. Zinc can be removed
and substituted by cobalt(II) without loss of activity. Cobalt(II) in CoCA is high
spin. Three histidine imidazoles and a water molecule are coordinated. One of
H
S
2 6,
H 2 His-1l9
H
H M
/ 0 N 6 N-H
H 62
H
N
H
N.
'H H
His-96
the imidazoles is coordinated to the metal through its N61 nitrogen, whereas the
other two are coordinated through their Ne2 nitrogens. If NO 3 is added to the
CoCA solution, it binds cobalt giving rise to a five-coordinate derivative where
one oxygen of nitrate adds as a fifth ligand. The spectrum is shown in Fig. 12-31
as an example of a spectrum of a metalloprotein. 1 17)
Four signals are observed far downfield. Three of them disappear when the
spectrum is recorded in D 2 0 and these are assigned to the exchangeable NH
protons of the three coordinated histidine rings. The non-exchangeable signal at
71 ppm is of similar line width, and is thus assigned to the H62 proton of His-1 19,
the only one not adjacent to a coordinating nitrogen. This signal gives an nOe
with the NH signal at 80 ppm, which is therefore its neighbor (Hs2) on the
imidazole ring of His- 119.(118) The other signals between + 30 and -20 ppm
belong to protons that are all close to cobalt and the shifts are all dipolar in
origin. Their assignment 1 18) is a challenge with implications of paramount
importance for further investigation of proteins.
I I I I I | | | | |
100 80 60 40 20 0 -20 -40
S(ppm)
FIGURE 12-31 Room temperature 1H nmr spectrum of the

nitrate adduct of cobalt(II) substituted carbonic
1 17 118
anhydrase.( ', The shaded signals disappear in D20.
12-15 THE INVESTIGATION OF BIMETALLIC SYSTEMS
We now consider the consequences of magnetic coupling in bimetallic systems

(see Chapter 11) on nmr. For simplicity and without loss of generality, consider
that one nucleus senses only one metal ion of the couple. If this were not the
case, the effects of each metal ion, calculated separately, would be added. Our
working scheme is therefore
N -- M M2
where N is the nucleus and M1 and M2 two metal ions. Also consider that
magnetic coupling between M1 and M2 is a weak perturbation with respect to
the energy and distribution of electrons of the moieties to which M, and M2
belong in the absence of magnetic coupling. This means that spin density on
nucleus N is not affected by the existence of magnetic coupling. Furthermore,
we can also consider that the electronic relaxation time of metal 1 changes its
value in a perturbative fashion when interacting with the electron of metal 2.
This consideration is not fully justified, since it is arbitrary to refer to only one
electronic relaxation time of a metal ion.* In principle, there is an electronic
relaxation time for each level arising from magnetic coupling. Nevertheless, this
reasoning leads to the conclusion that for homodimeric systems there is no change
* Electronic relaxation time is treated as a single transition; when there are several transitions
as in the case of zero field split levels, it is appropriate to assign a relaxation time for each transition.
of electronic relaxation times, whereas in heterodimetallic systems one metal ion

will have larger electronic relaxation times than the other. In the latter case, the
slow relaxing metal ion will take advantage of magnetic coupling, and its electrons
will relax with the mechanism of the fast-relaxing metal ion. Thus, magnetic
coupling causes a shortening of the electronic relaxation time of the slow-relaxing
metal ion. These conclusions are consistent with observations. Equations and
discussions are available in the literature that relate the extent and nature of J
3
with the values of electron relaxation times.' 119)
The 'H nmr of Dimeric Complexes

The following ligand
H H H H
CH3 H3C
H s 5 H
6
N N N N
H 2 H
3 3
indicated as PMK, provides a nice frame for bimetallic systems where antiferro-
magnetic coupling is established between the two metal ions. The systems studied
include CoCo, NiNi, CuCo, and CuNi.(' 20 ) With Zn being diamagnetic CuZn,
CoZn, and NiZn represent the blank.!70" 21 The J value is relatively small (i.e.,
<< kT at room temperature). Consequently, all the energy levels arising from
magnetic coupling are almost equally populated. The isotropic shift experienced
by each proton is the sum of the contributions of the two metal ions (i.e., the
TABLE 12-9. 200 MHz 'H nmr Shift and T, Values at 300 K for the Isotropically
Shifted Signals in the Cu2Co2SOD Derivative Together with Their Assignment
(signal labeling refers to Figs. 12-33 and 12-34)
Signal 5,ppm T1, ms Assignment
A 66.2 1.5 His-61 H62

B 56.5 7.8 His-118 H51
C 50.3 4.2 His-44 Hr2
D 49.4 3.8 His-69 H62
E 48.8 4.6 His-78 Ho2
F 46.7 2.1 His-78 Hc2 (His-69 He2)
G 40.6 3.5 His-44 HJ2
H 39.0 1.8 His-118 HEI
37.4 1.7 Asp-81 Hl1 (Asp-81 H#2)
J' 35.6 1.7 Asp-81 H#2 (Asp-81 H#1)
J 35.4 d His-69 Hr2 (His-78 Hr2)
K 34.5 8.0 His-46 H61
L 28.4 4.3 His-46 H62
M 25.3 2.7 His-44 HEl
N 24.1 2.9 His-1 18 H62
0 19.6 1.9 His-46 HEl
P 18.7 1.6 His-44 H#1
Q(R) -6.2 2.4 His-69 H#2
R(Q) -6.2 2.4 His-44 H#2
sum of the shift of a given proton in the MZn and the ZnM 2 complexes). As
far as the line widths are concerned, those of the CuZn complex are very large
for those protons sensing copper because , - 10-' sec and T, is given by r,
which is ~ 10-" sec. 2 0 ' Such a relatively long value of T, provides a short
nuclear T2 [equation (12-38)]. In the CuCo system, cobalt(II) has an electronic
relaxation time of ~ 10 -" sec (see Table 12-9) and the r, of copper approaches
this value, becoming much shorter than the value of the uncoupled system. The
H nmr signals are now quite sharp' 21 ) (Fig. 12-32).
-CH, 4-H
(Co) -CH3 (Cu)
10.5 ms (Cu 120 4-H
3 msms(o70 mns
3-H 5H5-H
(Co) 32Cms (CU)
24 ms 3ms3-H 40 ms
6-H (Co) 6-H (o
(CU) 26 ms
1.9 ms
*3.2 ms
J
140 120 100 80 60 40 20 0

FIGURE 12-32 1
H nmr spectrum of CuCo (PMK) 34+ in D20. Peaks of Co2(PMK) 34+
are indicated with an asterisk.0")
Cu2 Co2 Superoxide Dismutase

The enzyme Cu 2 Co 2 SOD, where SOD refers to superoxide dismutase, is
analogous to the bimetallic systems. SOD catalyzes the dismutation of the
superoxide ion. One common enzyme is a dimeric protein of MW equal to 32,000
with a MW of 16,000 in each subunit. In each subunit, zinc is bound to three
histidine nitrogens and one oxygen of an aspartate ( 22 ) (Fig. 12-33). Copper is
bound to four histidine nitrogens and is exposed to solvent (see Figure 12-33).
One histidine is in common to the two ions, bridging them. The protein rotates
with r, ~ 10-8 sec, whereas c, of copper(II) is 2 x 10-' sec.(65 ) The protons
sensing copper(II) have such a broad nmr line that they escape detection. The
protein can be manipulated to substitute zinc with cobalt. The copper and cobalt
ions experience antiferromagnetic coupling through the bridging imidazolate
with J = 17 cm- (123) The electronic relaxation times of copper(II) approach
those of cobalt(II), which are of the order of 10- 12 sec. The 'H nmr signals of
the ligands of both metal ions can now be observed since they are relatively
sharp.(4 8 "2 41 Through nOe it has been possible to propose a full assignment(l 1>
2
since several interproton distances could be calculated and compared to those

of the x-ray structure (see Table 12-9 and Fig. 12-34).
His-120
His-46
His-48
K
A Asp-83 D
His-63 His-71
ZnO
His-80
E
F,J
22
FIGURE 12-33 Scheme of the metal sites in Cu 2 Zn 2 SOD.(1 )
K
FIGURE 12-34 'H nmr B
spectrum at 300 K and pH 5.6 CDE GH MN
acetate buffer of the A
Cu2 Co2 SOD derivative. i J
F j, L OP 0,R3
I | I | | | I I I I I I I I
80 60 40 20 0 -20
1H nmr Spectra of Fe2S2 Ferredoxins

The Fe 2 S 2 ferredoxins are electron transfer proteins of low molecular weight
(~ 10,000) containing the following bimetallic moiety.
Cys-S S S-Cys
Fe Fe 7
Cy-- S S-Cys
12-15 The investigation of Bimetallic Systems 547
The oxidized form of the protein contains two iron(III) ions, whereas the reduced
form contains one iron(III) and one iron(II). In the oxidized form, the two iron(III)
ions are antiferromagnetically coupled with J = 400 cm-', giving an S' = 0
ground state. The energy levels, calculated according to equation (11-40) are
shown in Fig. 12--35. Since J ~k-T, only part of the levels are occupied and the
magnetic susceptibility of the system is reduced with respect to that expected for
an uncoupled system. The electronic relaxation times of iron(III) are long and
the 'H nmr signals of the f#-CH 2 (see the foregoing scheme) are expected to be
very broad. Although still very broad, they are observed in the dimer because
of the reduced magnetic susceptibility. All eight protons fall inside a broad
envelope (Fig. 12-36).
Fe(Ill) - Fe(Ill) Fe(lI) - Fe(lll)

S1 =5/2 S2=5/2 S1=2 S2= 5/2
6000 S'= 5
5J
FIGURE 12-35 Energy
levels for Fe2 S 2 dimetallic
4000 S' = 4 centers calculated for
J= 400 cm -1 in the oxidized
form and J= 200 cm-1 in the
reduced form. 1 26
S' = 9/2
S' = 3
2000 9/2J
S' = 7/2
3_
S' =2
7/2J
2J S'= 5/2
0
S'=
S'=0
1 fi 5/2J
=3/2
S'
S'/ = 1/2
HC-CH2 -S\ S /S-CH 2-CH

Fe Fe
CH2 -S S S-CH2 1
FIGURE 12-36 H nmr
HC CH spectra of oxidized and
reduced forms of Fe2S 2
Oxidized ferredoxin from
spinach.( 126 ,127) The signals
marked with asterisks belong
R cdA to impurities.
B D x4 E F H j
1 11 1 1 1 1 1 1 1 G I I I I I I
140 120 100 80 60 40 20 0 -20
S(ppm)
In the reduced case, iron(II) has short electronic relaxation times and the
whole system becomes more suitable for nmr investigation. The spectrum is
shown in Fig. 12-36.(127) The J value is now 200 cm-1 and the energy levels are
reported in Fig. 12-35. The ground state is given by S' = '/2 (i.e., by S' = S, - S2
where Si refers to iron(III) and S 2 to iron(II)). If only this level were populated,
the isotropic shifts of iron(III) would be far downfield (because <S,> is negative
according to equation (12-11) and the spin density for a delocalization is positive
[see equations (12-15) and (12-20)]. The isotropic shifts of iron(II) would be
upfield because of a negative S2 leading to a positive (S22>. The four signals far
downfield are assigned to the two fl-CH 2 protons of ligands bound to iron(III).
Figure 12-37. Signals F, G, H, and I are assigned to the #l-CH2 of ligands bound
to iron(II). They are slightly downfield because of population of excited S' levels.
The assignment is based on the temperature dependence of the shifts. The signals
assigned to iron(III) are expected to follow a Curie dependence (i.e., their shifts
decrease with increasing temperature because the contribution of iron(III) to the
magnetic susceptibility of the system decreases with increasing temperature as
seen qualitatively with the Curie law [see equation (12-12)]. In the case of
iron(II), the temperature increase causes an increase in the population of the
states with positive S., since the highest S' level is given by S, + S 2 . Accordingly,
the downfield isotropic shift increases with increasing temperature. In Fig. 12-37,
the experimental temperature dependence of the isotropic shifts is shown together
with expectations from considering the energy levels of Fig. 12-35.(126,127) These
results show that upon reduction of the oxidized protein only one localized iron
of the pair is reduced and a mixed valent system does not result.( 1 2 6) NOE studies
have suggested that the reducible iron is the one closer to the surface of the
2 8
protein. 1 >
140
A A
130-
B
120- B
110-
100 --
a. D
E 200Fe(III)
40- E
30- F F
G 0
20 - G
10 -200
3.3 3.5 3.7 3.3 3.5 3.7 0 2 4 6 8 10
1/T(K' X10 3 1IT(K'X10 3 )
(A) (B) (C)
FIGURE 12-37 Temperature dependence of the isotropic shifts of the 1H nmr
signals in reduced spinach (A) and P. umbilicalis (B) ferredoxin. The predicted
behavior is shown in (C).)126
Perspectives in Cluster Investigations

Paramagnetic polymetallic systems are becoming more readily available from
coordination chemistry and metalloproteins. Extension of the theory of dimetallic
systems(1 26 ) allows the investigation of more complex systems (e.g., Fe4 S4 , Fe3S4,
6 29
Co 4 S4 clusters). The reader is referred to specialized literature.(12 ,1 -133)
Shift Reagents
Rare earth tris-chelates of fl-diketonate derivatives are Lewis acids and cause
large pseudocontact shifts in Lewis bases that are added to solutions containing
these complexes. The resulting resonances for several of these ions (e.g., europium
and praseodymium complexes) are relatively sharp. In Fig. 12-38, the 100-MHz
proton magnetic resonance spectra of CDC13 solutions of cis-4-butylcyclohexanol
containing varying amounts of Eu(dpm) 3 [dpm is dipivaloylmethanato,
(CH 3)3CC(O)CHC(O)C(CH 3)3] are shown11 3 4 ) to illustrate this behavior. The
complex pattern obtained for the pure alcohol is shown in the bottom spectrum.
A first-order spectrum with all resonances well separated is obtained in curve D.
Note that the spin-spin couplings are still intact.
-(CHslhC-
Hr
I I FIGURE 12-38 100-MHz pmr
(E) -Flo spectra of cis-4-tert-

53.83 19.78 H(CHahC- butylcyclohexanol 4
OH H HAe Ha3. HH (20 mg, 1.28x 10 Mo) in
(D) .+10 various amounts of
30.0 HH
H, 0C2I4- Eu(dPM) 3 : (A) 0.0 mg; (B)
(C) H
---- g C)1 .0m ;D
H+10 [Reprinted with permission
H Hi Haa4II from P. V. Demarco et al., J.
I Am. Chem. Soc., 92, 5734
(B)H (B) 2 3(1970). Copyright by the
HC#1+1H +10 American Chemical Society.]
OH H
(A) +10
1 110 0.0 910 8.0' 710 1F.0 4.02 31.00 .0 to GU0
Rare earth complexes that exhibit this behavior are called"")5 shift reagents,
S.R. As is usually the case with these systems, the S.R. complex is in fast exchange
with an excess of the Lewis base. The lanthanide complexes of the fluorinated
chelate 1,1,1b,2,2,3,3-heptafluoro-7,7-dimethyl-4,6-octanedione, (CH 3 )3 CucO)
CHC(O)CFCF 2CF 3 , abbreviated as fod, are stronger Lewis acids and produce
larger shifts. (136 Since these early reports there have been hundreds of papers v
dealing with applications of shift reagents. (137) High field spectrometers have
caused a decrease in interest in shift reagents. Here, we shall deal with examples
of some different types of applications and discuss some of the potential

complications encountered.
One use of a shift reagent involves simplifying second-order (i.e., J ; A)
spectra. Different protons in the molecule are shifted by different amounts upon
complexation. Thus, as the shift reagent concentration is increased, A becomes
greater than J. This behavior is seen in Fig. 12-38.
A second potentially exciting application of shift reagents involves their use
in determining geometries of molecules in solution." 3 8 This experiment is usually
done in the fast exchange region. The proton nmr spectral shifts induced by shift
reagents are assumed to be almost exclusively dipolar in origin. Otherwise,
factorization is needed. Ideally, a structure would be assumed for the molecule
and equation (12-23) would be used to calculate the expected dipolar shifts at a
large number of various nuclei in the molecule whose structure is to be determined.
The assumed structure of the molecule would be varied to produce a best fit of
the experimental shifts. Since the structure of the molecule being investigated and
that of the complex in solution are not known, nor are the magnitude and position
of the magnetic dipole of the metal center in the complex, there are eight unknowns
in the system. These unknowns are best seen by looking at Fig. 12-39, where the
example of a rigid ligand such as pyridine complexed to a S.R. is illustrated. Four
parameters are needed to fix the orientation of the molecule being studied relative
to the shift reagent: (1) the metal-donor atom distance, r; (2) the ligand
atom-donor atom-metal bond angle, a; (3) #l, the angle between the ligand's
molecular plane and the magnetic xy-plane of the metal; and (4) y, the angle
FIGURE 12-39 Definition of

the variables in a shift r
reagent determination of
molecular structure.
measuring the twist of the ligand's molecular plane around the nitrogen-
para-carbon axis. In addition, two angles are needed to define the orientation of
the magnetic axis with respect to the metal-donor bond. Two additional unknowns
are needed to account for the anisotropy in X.To complete our discussion of a
rigorous solution of this problem, we should consider the possibility of having
significant amounts of both 2: 1 and 1: 1 complexes in solution. (Most shift
reagents are potentially diacidic, having two acid sites.) If neither the 2:1 nor
the 1:1 complex is the predominant species in solution, both K, and K must
2
be known as well as two sets of the unknowns mentioned before, one set for each
of the two complexes. However, it is obvious from the preceding discussion that
it is not simple to determine geometries of molecules in solution in a rigorous
manner. Thus, to gain any information about geometries in solution, more
independent information is needed, such as the relaxation rates, and some
simplifying assumptions must be made.
If threefold or higher axial symmetry and the presence of only one
predominant species are assumed, the problem becomes tractable. With axial
symmetry, equation (12-25) can be used instead of equation (12-25). By taking
the ratio of all the observed nmr shifts with respect to the largest observed shift,
Avj, equation (12-53) can be derived from equation (12-25):
Av (1-3 cos 2O,'\ (1 3 cos2'0)

Av; r,Av-
i33 - -I
r i,3(12-53)
and the magnitude of the magnetic anisotropy in the complex is eliminated from
the problem. (This is possible only if one ignores the chelate ligand protons and
works only with the coordinated Lewis base.)
The complications to obtain structures in solution include:
1. The stoichiometry of the adduct in solution must be determined but is

often assumed.
2. The treatment assumes that there is only a single geometrical isomer of
the adduct in solution. Many are possible. Furthermore, lanthanides have a high
affinity for water, which if present in the system can compete for Lewis acid sites
with the substrate, producing even more complexes in the system.
3. When the substrate is involved in a dynamic, intramolecular process,
the average conformation of the molecule is determined, and this might represent
a relatively unstable one.
4. The complex is generally assumed to be axially symmetric, so that the
dipolar shifts can be evaluated with a <(1 - 3 cos2 O)/r3 > term. Those systems
studied to date do not possess axial symmetry (however, see reference 139).*
5. The location of the principal magnetic axis in the complex is often not
known relative to the substrate ligand.
6. When the substrate is a complex one, it is necessary to assume that there
is only one binding site.
*Some more accurate studies were possible in proteins where many experimental data are
available, and in this case no assumption of this type is needed."-"
In spite of these complications, the method is capable of distinguishing gross

structural features such as those that exist in many different isomeric compounds.
For example, the compounds
34
are readily distinguished,' ) as are (41)
C6 H1
Shift reagents have been used to differentiate between internal and external
phospholipid layers in membranes. 1 4 2 They have also been employed in
conformational studies of nucleotides in solution(1 43 ) and in other metalloprotein
systems.( 4 4 ) one must always be concerned with the problem that coordination
of the shift reagent may affect the conformation of the biomolecule.
Another interesting application involves the use of optically active shift
reagents to determine optical purity. The idea is similar to that discussed in
Chapter 8, in which optically active solvents were employed. Here, different
stability constants for forming the different diastereoisomeric adducts exist,
leading to different mole fraction averaged shifts for the enantiomeric bases.
Several reports of different optically active rare earth complexes that can be used
45
for this application have appeared." -147)
REFERENCES 1. R. J. Fitzgerald and R. S. Drago, J. Am. Chem. Soc., 90, 2523 (1968).
CITED 2. I. Bertini and C. Luchinat, "NMR of Paramagnetic Molecules in Biological Systems,"
Benjamin/Cummings, Menlo Park, CA, 1986.
3. J. S. Griffith, "The Theory of Transition Metal Ions," Cambridge University Press,
London, 1971.
4. R. M. Golding, "Applied Wave Mechanics," Van Nostrand, London, 1969.
5. E. Fermi, Z. Phys., 60, 320 (1930).
6. H. M. McConnell, Proc. Nati. Acad. Sci. USA, 69, 335 (1972).
7. R. J. Kurland and B. R. McGarvey, J. Magn. Reson., 2, 286 (1970).
8. W. deW. Horrocks, Jr. and D. D. Hall, Inorg. Chem., 10, 2368 (1971).
9. R. M. Golding, R. 0. Pascual, and B. R. McGarvey, J. Magn. Reson., 46, 30 (1982).
10. R. M. Golding, R. 0. Pascual, and J. Vrbancich, Mol. Phys., 31, 731 (1976).
11. B. Bleaney, J. Magn. Reson., 8, 91 (1972).
12. R. M. Golding and M. P. Halton, Aust. J. Chem., 25, 2577 (1972).
13. C. N. Reilley, B. W. Good, and J. F. Desreux, Anal. Chem., 47, 2110 (1975).
14. J. Reuben and G. A. Elgavish, J. Magn. Reson., 39, 421 (1980).
15. B. Bleaney, C. M. Dobson, B. A. Levine, R. B. Martin, R. J. P. Williams, and A. V.
Xavier, J. Chem. Soc. Chem. Commun., 791 (1972).
16. J. G. Shelling, M. E. Bjornson, R. S. Hodges, A. K. Taneja, and B. D. Sykes, J. Magn.
Reson., 57, 99 (1984).
17. H. M. Goff, J. Am. Chem. Soc., 103, 3714 (1981).
18. G. N. LaMar, G. R. Eaton, R. H. Holm, and F. A. Walker, J. Am. Chem. Soc., 95,
63 (1973).
19. W. DeW. Horrocks, Inorg. Chem., 9, 690 (1970).
20. C. Benelli, 1. Bertini, and D. Gatteschi, J. Chem. Soc. Dalton, 661 (1972).
21. D. Doddrell and J. D. Roberts, J. Am. Chem. Soc., 92, 6839 (1970).
22. W. DeW. Horrocks and D. D. Hall, Inorg. Chem., 10, 2368 (1971).
23. M. Gerloch and R. C. Slade, J. Chem. Soc., A, 1022 (1969).
24. R. H. Holm, G. W. Everett, and A. Chakravorty, Progr. Inorg. Chem., 7, 83 (1966).
25. W. DeW. Horrocks and E. S. Greenberg, Inorg. Chem., 10, 2190 (1971).
26. E. A. Lalancette and D. R. Eaton, J. Am. Chem. Soc., 86, 5145 (1964).
27. J. P. Jesson, J. Chem. Phys., 45, 1049 (1966).
28. F. Keffer, T. Oguchi, W. O'Sullivan, and J. Yamashita, Phys. Rev., 115, 1553 (1959).
29. W. D. Perry, R. S. Drago, D. W. Herlocker, G. K. Pagenkopf, and K. Czworniak,
Inorg. Chem., 10, 1087 (1971).
30. J. P. Quaegebeur, C. Chachaty, and T. Yasukawa, Mol. Phys., 37, 409 (1979).
31. H. M. McConnell, Proc. Natl. Acad. Sci. USA, 69, 335 (1972).
32. E. W. Stone and A. H. Maki, J. Chem. Phys., 37, 1326 (1962).
33. M. Karplus and G. K. Fraenkel, J. Chem. Phys., 35, 1312 (1961).
34. E. T. Strom, G. R. Underwood, and D. Jurkowitz, Mol. Phys., 24, 901 (1972); W.
Derbishire, Mol. Phys., 5, 225 (1962).
35. D. R. Eaton, A. D. Josey, W. D. Phillips, and R. E. Benson, Mol. Phys., 5,407 (1962).
36. "NMR of Paramagnetic Molecules," eds. G. N. LaMar, W. DeW. Horrocks, Jr., and
R. H. Holm, Academic Press, New York, 1973.
37. R. W. Kluiber and W. DeW. Horrocks, Inorg. Chem., 6, 430 (1967).
38. I. Bertini and L. J. Wilson, J. Coord. Chem., 1, 237 (1971).
39. R. H. Holm, A. Chakravorty, and G. 0. Dudek, J. Am. Chem. Soc., 86, 379 (1964).
40. a. M. Wicholas and R. S. Drago, J. Am. Chem. Soc., 91, 5463 (1969).
b. I. Bertini, D. Gatteschi, and A. Scozzafava, Inorg. Chim. Acta, 6, 185 (1972).
41. I. Bertini, C. Luchinat, and A. Scozzafava, Inorg. Chim. Acta, 19, 201 (1976).
42. G. N. LaMar, W. DeW. Horrocks, and L. C. Allen, J. Chem. Phys., 41, 2126 (1964).
43. J. A. Happe and R. L. Ward, J. Chem. Phys., 39, 1211 (1963).
44. R. W. Kluiber and W. DeW. Horrocks, Inorg. Chem., 6, 166 (1967).
45. R. H. Holm, G. W. Everett, and W. DeW. Horrocks, J. Am. Chem. Soc., 88, 1071(1966).
46. a. G. M. Zhidomirov, P. V. Schastnev, and N. D. Chuvylkin, J. Struct. Chem., 11,
458 (1970).
b. R. E. Cramer and R. S. Drago, J. Am. Chem. Soc., 92, 66 (1970).
47. I. Bertini, G. Canti, C. Luchinat, and F. Mani, J. Am. Chem. Soc., 103, 7784 (1981).
48. L. Banci, 1. Bertini, C. Luchinat, and A. Scozzafava, J. Am. Chem. Soc., 109, 2328
(1987).
49. I. Bertini, C. Luchinat, and A. Scozzafava, Inorg. Nucl. Chem. Lett., 15, 89 (1979).
50. 1. Bertini, D. L. Johnston, and W. DeW. Horrocks, Inorg. Chem., 9, 693 (1970); ibid.,
p. 698.
51. R. J. Fitzgerald and R. S. Drago, J. Am. Chem. Soc., 89, 287 (1967).
52. I. Morishima, T. Yonezawa, and K. Goto, J. Am. Chem. Soc., 92, 6651 (1970).
53. L. Banci, I. Bertini, and C. Luchinat, "Electron and Nuclear Relaxation," VCH,
Heidelberg, 1991.
54. L. Nordenskjold, A. Laaksonen, and J. Kowalewski, J. Am. Chem. Soc., 104, 379
(1982).
55. D. Waysbort and G. Navon, J. Chem. Phys., 68, 3704 (1978).
56. H. P. W. Gottlieb, M. Barfield, and D. M. Doddrell, J. Chem. Phys., 67, 3785
(1977).
57. D. M. Doddrell, P. C. Healy, and M. R. Bendall, J. Magn. Reson., 29, 163 (1978).
58. N. Bloembergen, J. Chem. Phys., 27, 575 (1957).
59. 1. Solomon, Phys. Rev., 99, 559 (1955).
60. M. Gueron, J. Magn. Reson., 19, 58 (1975).
61. A. J. Vega and D. Fiat, Mol. Phys., 31, 347 (1976).
62. H. Sternlicht, J. Chem. Phys., 42, 2250 (1965).
63. 1. Bertini, C. Luchinat, M. Mancini, and G. Spina, in "Magneto-Structural Corre-
lations in Exchange Coupled Systems", eds. D. Gatteschi, 0. Kahn, and R. D. Willett,
Reidel, Dordrecht, 1985, p. 421.
64. I. Bertini, C. Luchinat, M. Mancini, and G. Spina, J. Magn. Reson., 59, 213 (1984).
65. I. Bertini, F. Briganti, C. Luchinat, M. Mancini, and G. Spina, J. Magn. Reson., 63,
41 (1985).
66. 1. Bertini, G. Lanini, C. Luchinat, M. Mancini, and G. Spina, J. Magn. Reson., 63,
56 (1985).
67. L. Banci, I. Bertini, and C. Luchinat, Inorg. Chim. Acta, 100, 173 (1985).
68. I. Bertini, C. Luchinat, and J. Kowalewski, J. Magn. Reson., 62, 235 (1985).
69. L. Banci, I. Bertini, F. Briganti, and C. Luchinat, J. Magn. Reson., 66, 58 (1986).
70. C. Owens, R. S. Drago, I. Bertini, C. Luchinat, and L. Banci, J. Am. Chem. Soc.,
108, 3298 (1986).
71. I. Bertini, C. Luchinat, C. Owens, and R. S. Drago, J. Am. Chem. Soc., 109,5208 (1987).
72. W. G. Espersen and R. B. Martin, J. Am. Chem. Soc., 98, 40 (1976).
73. A. S. Anderson and A. G. Redfield, Phys. Rev., 116, 583 (1959).
74. S. H. Koenig and R. D. Brown, III, in "NMR Spectroscopy of Cells and Organisms,"
CRC Press, Boca Raton, FL 1987.
75. H. S. Gutowsky, D. M. McCall, and C. P. Slichter, J. Chem. Phys., 21, 279 (1953).
76. H. S. Gutowsky and C. H. Holm, J. Chem. Phys., 25, 1228 (1956).
77. T. J. Swift and R. E. Connick, J. Chem. Phys., 37, 307 (1962).
78. R. A. Dwek, "Nuclear Magnetic Resonance in Biochemistry," Oxford University
Press, Oxford, 1973, Chapter 11.
79. R. Hausser and F. Noack, Z. Physik, 182, 93 (1964).
80. B. P. Gaber, R. D. Brown, S. H. Koenig, and J. A. Fee, Biochim. Biophys. Acta, 271,
1 (1972).
81. L. T. Muus and P. W. Atkins, eds., "Electron Spin Relaxation in Liquids," Plenum
Press, New York and London, 1972.
82. N. W. Ashcroft and N. D. Mermin, "Solid State Physics," Saunders, Philadelphia,
1976.
83. A. A. Manenkov and R. Orbach, "Spin Lattice Relaxation in Ionic Solids," Harper
& Row, New York, 1966.
84. R. Orbach, Proc. R. Soc. A, 264, 458 (1961); ibid., p. 485.
85. W. B. Lewis and L. 0. Morgan, Trans. Met. Chem., 4, 33 (1968).
86. J. H. Van Vleck, Phys. Rev., 57, 426 (1940).
87. M. Maekawa, S. Kitagawa, M. Munkata, and H. Masuda, Inorg. Chem., 28, 1904
(1989).
88. S. H. Koenig and R. D. Brown, III, in "Metal Ions in Biological Systems", eds.
H. Sigel and A. Sigel, Vol. 21, Dekker, New York, 1987.
89. I. Bertinin, F. Capozzi, and C. Luchinat, Magn. Reson. Imaging, 9, (1991); S. H.
Koenig and R. D. Brown III, Prog NMR Spectr 22, 487 (1990).
90. G. Hernandez, M. F. Tweedle, and R. G. Bryant, Inorg. Chem., 29, 5109 (1990).
91. J. H. Noggle and R. E. Schirmer, "The Nuclear Overhauser Effect," Academic Press,
New York, 1971.
92. D. Neuhaus and M. Williamson, "The Nuclear Overhauser Effect in Structural and
Conformational Analysis," VCH, New York, 1989.
93. L. B. Dugad, G. N. LaMar, L. Banci, and I. Bertini, Biochemistry, 29, 2263 (1990).
94. L. Banci, A. Bencini, I. Bertini, C. Luchinat, and M. Piccioli, Inorg. Chem., 29, 4867
(1990).
95. 1. Bertini, L. Banci, C. Luchinat, and M. Piccioli, FEBS Lett., 272, 175 (1990).
96. C. Luchinat, S. Steuernagel, and P. Turano, Inorg. Chem., 29, 4351 (1990).
97. S. D. Emerson and G. N. LaMar, Biochemistry, 29, 1545 (1990).
98. Y. Fengo and S. W. Englander, Biochemistry, 29, 3505 (1990).
99. N. C. Veitch and R. J. P. Williams, Eur. J. Biochem., 189, 351 (1990).
100. W. D. Phillips and R. E. Benson, J. Chem. Phys., 33, 607 (1960); R. E. Benson, D.
R. Eaton, A. D. Josey, and W. D. Phillips, J. Am. Chem. Soc., 83, 3714 (1961).
101. D. R. Eaton and W. R. McClellan, Inorg. Chem., 6, 2134 (1967).
102. D. R. Eaton, A. D. Josey, and R. E. Benson, J. Am. Chem. Soc., 89, 4040 (1967).
103. D. R. Eaton, A. D. Josey, W. D. Phillips, and R. E. Benson, J. Chem. Phys, 37, 347
(1962).
104. W. DeW. Horrocks, Inorg. Chem., 9, 690 (1970).
105. R. E. Ernst, M. J. O'Connor, and R. H. Holm, J. Am. Chem. Soc., 90, 5735 (1968).
106. D. H. Gerloch and R. H. Holm, J. Am. Chem. Soc., 91, 3457 (1969).
107. L. J. Wilson and I. Bertini, J. Coord. Chem., 1, 237 (1971).
108. G. N. LaMar and L. Sacconi, J. Am. Chem. Soc., 89, 2282 (1967).
109. 1. Bertini, L. Sacconi, and P. Speroni, Inorg. Chem., 11, 1323 (1972).
110. a. I. M. Walker and R. S. Drago, J. Am. Chem. Soc., 90, 6851 (1968).
b. D. G. Brown and R. S. Drago, J. Am. Chem. Soc., 92, 1871 (1970).
c. Y. Y. Lim and R. S. Drago, J. Am. Chem. Soc., 94, 84 (1972).
111. 1. Bertini, C. Luchinat, and E. Borghi, Inorg. Chem., 20, 303 (1981).
112. L.-P. Hwang and J. H. Freed, J. Chem. Phys., 63, 4017 (1975); J. H. Freed, J. Chem.
Phys., 68, 4034 (1978).
113. C. F. Polnaszek and R. G. Bryant, J. Chem. Phys., 81, 4038 (1984).
114. C. C. Lester and R. G. Bryant, J. Phys. Chem., 94, 3654 (1990).
115. I. Bertini, F. Capozzi, C. Luchinat, and P. Turano, J. Magn. Reson., (1991), in press.
116. G. N. LaMar and F. A. Walker, in "The Porphyrins," ed. D. Dolphin, Academic
Press, New York, 1979, Vol. 4, p. 61.
117. I. Bertini, G. Canti, C. Luchinat, and F. Mani, J. Am. Chem. Soc., 103, 7784 (1981).
118. Unpublished results from our laboratory.
119. I. Bertini, G. Lanini, C. Luchinat, M. Mancini, and G. Spina, J. Magn. Reson., 63,
56 (1985).
120. C. Benelli, A. Dei, and D. Gatteschi, Inorg. Chem., 21, 1284 (1982).
121. I. Bertini, C. Owens, C. Luchinat, and R. S. Drago, J.Am. Chem. Soc., 109,5208 (1987).
122. J. A. Tainer, E. D. Getzoff, K. M. Beem, J. S. Richardson, and D. C. Richardson, J.
Mol. Biol., 160, 181 (1982).
123. I. Morgenstern-Badarau, D. Cocco, A. Desideri, G. Rotilio, J. Jordanov, and N.
Dupre, J. Am. Chem. Soc., 108, 300 (1986).
124. I. Bertini, G. Lanini, C. Luchinat, L. Messori, R. Monnanni, and A. Scozzafava, J.
Am. Chem. Soc., 107, 4391 (1985).
125. L. Banci, I. Bertini, C. Luchinat, M. Piccioli, A. Scozzafava, and P. Turano, Inorg.
Chem., 28, 4650 (1989).
126. L. Banci, 1. Bertini, and C. Luchinat, Struct. Bonding, 72, 113 (1990).
127. 1. Bertini, G. Lanini, and C. Luchinat, Inorg. Chem., 23, 2729 (1984).
128. L. B. Dugad, G. N. LaMar, L. Banci, and 1. Bertini, Biochemistry, 29, 2263 (1990).
129. I. Bertini, F. Briganti, C. Luchinat, A. Scozzafava, and M. Sola, J. AM. Chem. Soc.,
113, 1237 (1991).
130. I. Bertini, F. Briganti, and C. Luchinat, Inorg. Chim. Acta, 175, 9 (1990).
131. I. Bertini, C. Luchinat, L. Messori, and M. Vasak, J. Am. Chem. Soc., 111, 7296
(1988); ibid, p. 7300.
132. L. Banci, . Bertini, F. Briganti, and C. Luchinat, New J. Chem., 15, 467 (1991).
133. L. Banci, I. Bertini, F. Briganti, C. Luchinat, A. Scozzafava, and M. Vicens Oliver,
Inorg. Chem., (1991), in press.
134. P. V. Demarco et al., J. Am. Chem. Soc., 92, 5734 (1970).
135. C. C. Hinckley, J. Am. Chem. Soc., 91, 5160 (1969).
136. R. E. Rondeau and R. E. Sievers, J. Am. Chem. Soc., 93, 1522 (1971).
137. G. E. Hawkes et al., "Nuclear Magnetic Shift Reagents," ed. R. E. Sievers, Academic
Press, New York (1973).
138. M. F. Rettig and R. S. Drago, J. Am. Chem. Soc., 88, 2966 (1966).
139. G. N. LaMar and E. A. Metz, J. Am. Chem. Soc., 96, 5611 (1974).
140. L. Lee and B. D. Sykes, Biochemistry, 22, 4366 (1983).
141. T. Akutani et al., Tetrahedron Lett., 1115 (1971).
142. S. B. Andrews et al., Proc. Nat. Acad. Sci. USA, 70, 1814 (1973) and references therein.
143. C. D. Barry et al., Biochim. Biophys. Acta, 262, 101 (1972) and references therein.
144. R. A. Dwek et al., J. Biochem., 21, 204 (1971).
145. G. M. Whitesides and D. W. Lewis, J. Amer. Chem. Soc., 92, 6979 (1970); 93, 5914
(1971).
146. H. L. Goering et al., J. Am. Chem. Soc., 93, 5913 (1971).
147. R. R. Fraser et al., Chem. Commun., 1450 (1971).
EXERCISES 1.For each of the following, state whether the contact contribution to the isotropic shift
in the proton nmr is close to zero. Also state whether the pseudocontact contribution
is zero. Explain all answers.
a. Ni(: N 6 2+ (octahedral about Ni).
b. The n-butanol adduct of Eu(dpm) 3 , where dpm is
O 0 O
I!
(CH 3 ) 3C C C(CH 3 )3
H
2. The following Av(iso) values are reported for Ni(CHCH 2 NH2)2+: NH 2, -105
CH 2, 34.7; phenyl protons: o, -1.3; m, 1.6; p, -1.4 (ppm). Interpret these shifts in
terms of delocalization mechanisms.
3. The stable free radical illustrated gives a sharp three-line esr signal in benzene solution
at room temperature with a concentration of about 10-4 M. The nmr signal is too
broad to be detected. On increasing radical concentration, the esr signal broadens
and is barely discernible at the 1 M level, though now the following nmr signal is
seen at 60 MHz.
Exercies 557
12 H CH
Relative H eH3
intensity: 2 4H
H
H CH 3
TMS H CH, 3
A
2000 Hz -
a. Explain the low concentration esr signal and the high concentration nmr signals.
b. Why does the esr signal disappear with increasing concentration while the nmr
signal grows with concentration?
4. Four-coordinate complexes of Ni(II) may be (1) low spin (S = 0) square planar, (2)
high spin (S = 1) tetrahedral, (3) an intermediate D2 geometry with singlet and triplet
state close in energy, or (4) situations (1) and (2) in equilibrium and interchanging
with a rate constant on the order of 107 sec- .
a. Describe what result you would expect to obtain for each of the above cases if the
magnetism were studied as a function of temperature (i.e., Curie-Weiss behavior
or not, and why).
b. Could cases (3) and (4) be distinguished by employing infrared, nmr, or esr
spectroscopy? (Indicate what is expected, and why, for each of those techniques.)
c. Describe what you would expect to see from a study of the nmr spectrum and its
temperature dependence for each of the above four cases.
5. Water has one - MO capable of t-bonding with the t,, orbitals in octahedral hexaqua
complexes. Predict the order of 'H hyperfine shifts in the d'-d' series.
6. Oxygen-17 nmr has been observed in aqueous solutions of paramagnetic ions. Water
that is coordinated to a paramagnetic ion has an 170 resonance that is shifted far
up- or downfield in relation to uncoordinated water. An aqueous solution containing
the paramagnetic ion Dy 3 has only one 17 0 resonance. Aqueous solutions of the
diamagnetic ion Al 3 + exhibit two 17 0 resonances. See diagram of spectra.
a. Briefly explain why one 170 resonance is observed in aqueous Dy 3 + solutions,

while two are observed in Al 3 solutions.
b. A solution containing 0.1 mole of Dy 3 + and 10.0 moles of water has an 170
resonance 190 ppm upfield from uncoordinated water. Addition of 0.2 mole of
anhydrous Al 3 + causes this peak to shift to 216 ppm. Explain.
c. Calculate the number of water molecules in the coordination sphere of Al 3 .
Dy 3 +, H 20
Al 3 +, H2 0
A13+ H20, Dy 3+
0 ppm 200 ppm
7. Calculate the contact shift of a proton in a metal complex of axial symmetry if (a) the
isotropic shift is 103 ppm, (b) the metal coordinates are x = 0, y = 0, z = 0, and the
proton coordinates are x = 3.51, y = 2.76, z = 1.12 A, (c) the magnetic susceptibility
anisotropy is 2.3 x 10 8 m3 mol-', (d) the direction cosines of the z-axis of the
y-tensor are a = 0.128, b = 0.692, c = 0.710.
8. Are T, and T, of a proton at 300 MHz mainly contact, mainly dipolar, or both contact
and dipolar in origin if the contact shift is + 173 ppm, the metal-proton distance is
4.20 A, r, is 7 x 10- " sec and the electron relaxation time is
a. 10-12 sec?
b. 10-8 sec?
9. Determine the magnetic field at which the line width, expressed both in Hz and ppm,
is sharpest for a proton at 5.2 A from a metal with five unpaired electrons in a complex
having r, = 3 x 10-8 sec and T, = 10-" sec. Neglect the contact contribution.
10. Predict and explain the order of the electron relaxation times for the following metal
ions: Cu 2 +, Co 2 t (H.S.), Yb3+.
11. Predict the relative ease of observing COSY and NOESY cross peaks at 200 and 600
MHz between two signals in the paramagnetic complex of exercise 9. Remember that
during the buildup of the cross correlation, the COSY intensities decrease with T
2
and the NOESY intensities decrease with T,.
12. Calculate the distance between two protons A and B in a paramagnetic complex with
, = 10- sec knowing that, by irradiating A in a steady-state experiment at 500 MHz,
a nOe of 2.5% on B is obtained and the T, of B is 25 msec.
13. Explain why 'H nmr signals from the metal ligands in a antiferromagnetically coupled
nickel(II)-copper(II) system are sharper than in the analogous zinc(II)-copper(II)
system.
14. Are the nmr lines of a copper(II) complex broadest
a. in the presence of antiferromagnetic coupling with a high spin cobalt(II) complex?
b. in the presence of ferromagnetic coupling of equal magnitude?
c. in the absence of magnetic coupling?

Electron Paramagnetic
Resonance Spectra of 13
Transition Metal
Ion Complexes
13-1 INTRODUCTION
This chapter is an extension of Chapter 9; the principles covered there, as well

as those in Chapters 10 and 11, should be well understood before this chapter
is begun. It is for this reason that the presentation of the subject matter in this
chapter did not follow Chapter 9 directly.
The epr spectra of transition metal ion complexes contain a wealth of
information about their electronic structures. The additional information and
accompanying complications that are characteristic of transition metal ion
systems arise because of the approximate degeneracy of the d-orbitals and because
many of the molecules contain more than one unpaired electron. These properties
give rise to orbital contributions and zero-field effects. As a result of appreciable
orbital angular moments, the g-values for many metal complexes are very
anisotropic. Spin-orbit coupling also gives rise to large zero-field splittings (of 10
cm or more) by mixing ground and excited states.
An important theorem that summarizes the properties of multielectron
systems is Kramers' rule.") This rule states that if an ion has an odd number of
electrons, the degeneracy of every level must remain at least twofold in the absence
of a magnetic field. With an odd number of electrons, m, quantum numbers will
be given by ± 2 to + J. Therefore, any ion with an odd number of electrons
must always have as its lowest level at least a doublet, called a Kramers' doublet.
This degeneracy can then be removed by a magnetic field, and an epr spectrum
should be observed. On the other hand, for a system with an even number of
electrons, i, = 0, i , ... +J. The degeneracy may be completely removed by
a low symmetry crystal field, so only singlet levels remain that could be separated
by energies so large that an epr transition would not be observed in the microwave
region. This discussion is illustrated by the energy level splittings in Fig. 13-1.
For the even electron system, the ground state is non-degenerate and the J = 0
to 1 transition energy is quite frequently outside the microwave region.
A few additional comments relative to some experimental procedures (others
are given in Chapter 9) are in order in the way of introduction to transition metal
systems. A number of factors, other than those that are instrumental, affect the
epr line width. As in nmr, spin-lattice, spin-spin, and exchange interactions are
important.
Broadening due to spin-lattice relaxation results from the interaction of the 559
560 Chapter 13 Electron Paramagnetic Resonance Spectra of Transition Metal Ion Complexes
3
'A2 p 3T__
3
A mj
T2J (3)
5/2 J 2
(6)5/2 2 T, 2.< 0
( 3/2 1 0
() 2 1/2 2 _ 1
-1/2 3 T
(2)1/2 0
Gaseous Oh Spin-orbit Magnetic Gaseous Oh Spin-Orbit Magnetic

ion Crystal Coupling Field Ion Crystal Coupling Field
Field (Zero-field) Field (Zero-field) H
(A) (B)
FIGURE 13-1 The splitting of the gaseous ion degeneracy of (A) 002+ and (B) V+
by the crystal field, spin-orbit coupling, and a magnetic field. The T state is
regarded as having an effective L, called L, equal to 1; for an A state, L' = 0. Then
J = L' + S, .. . , L' - S. Only the zero-field and magnetic field splittings of the
ground state are shown.
paramagnetic ions with the thermal vibrations of the lattice. The variation in
spin-lattice relaxation times in different systems is quite large. For some
compounds the lifetime is sufficiently long to allow the observation of spectra at
room temperature, while in others this is not possible. Since these relaxation
times generally increase as the temperature decreases, many of the transition
metal compounds need to be cooled to liquid N2 or liquid He temperatures
before well-resolved spectra are observed.
Spin-spin interaction results from the magnetic fields that originate in
neighboring paramagnetic ions. As a result of these fields, the total field at each
ion is slightly altered and the energy levels are shifted. A distribution of energies
results, which produces a broad signal. This effect varies as (1/r3 ) -(1 - 3 cos 2 0),
where r is the distance between ions and 0 is the angle between the field and the
symmetry axis. Since this effect is reduced by increasing the distance between
paramagnetic ions, it is often convenient to examine transition metal ion systems
by diluting them in an isomorphous diamagnetic host. For example, a copper
complex could be studied as a powder or single crystal by diluting it in a host
lattice of an analagous zinc complex or by examining it in a frozen solution.
Dilution of the solid isolates the electron spin of a given complex from that of
another paramagnetic molecule, and the spin lifetime is lengthened. (Recall our
discussion of the spectra in Fig. 9-25.) If a frozen solution is used, it must form
a good glass; otherwise, paramagnetic aggregates form, which lead to a spectrum
with broadened lines. It is often necessary to remove 02 from the solvent because
this can lead to a broadening of the resonance. Even in a well-formed glass, one
cannot usually detect hyperfine splittings smaller than 3 or 4 gauss.
Line widths are altered considerably by chemical exchange processes. This
effect can also be reduced by dilution. If the exchange occurs between equivalent
paramagnetic species, the lines broaden at the base and become narrower at the
center. When exchange involves dissimilar ions, the resonances of the separate
lines merge to produce a single line, which may be broad or narrow depending
upon the exchange rate. Such an effect is observed for CuSO, 5H20 which has
two distinct copper sites per unit cell."'
In single crystals the anisotropy in the esr parameters can be obtained.
Information about the anisotropy in the system can also often obtained in powders
and glasses, because the resulting spectra are not those of a motionally averaged
system. We will next consider why anisotropic information is obtained from the
spectra even though the molecules in a glass or powder sample exist in an
extremely large number of possible orientations relative to the applied field.
Consider a molecule with a threefold or higher symmetry axis, which can be
described by a g, and a g,. As can be seen from Fig. 13-2, there are many axes
g,
g,
FIGURE 13-2 The g- and

g some g-axes in a crystal
with a threefold or higher
axis.
that could be labeled gi. In a bulk sample containing many orientations for the
assemblage of crystallites, there are more possible orientations that have the gi
axes aligned with the applied field than there are orientations that have the gl
axis aligned. The g-value for any orientation is given by:
2
g2 = g2 sin 2 + g cos (13-1)
where 0 is the angle that the principal axis (i.e., the gl axis) makes with the applied
field. Since all orientations of the crystallites in a solid are equally probable,
absorption will occur at all fields between that associated with gi and that
associated with g,. Since many more crystallites have g± aligned than g , the
most intense absorption will correspond to g,. When one considers the
probabilities of the various orientations and the transition probability corre-
FIGURE 13-3 idealized

absorption (A) and derivative
(B) spectra for an unoriented
system with S = '/2, axial
symmetry, and no hyperfine
interaction (gi > gg). 9
H inc. H inc_.
(A) (B)
FIGURE 13-4 Powder epr

spectra of S = 1/2 systems.
(A) An orthorhombic system
with / = 0; (B) isotropic g
with / = 1/2, g1 > g1, and
Al > A1 . In the latter cases,
accurate g- and A-values are
available only from computer
simulation.
AA
g, |I
Ai
l' AlA
sponding to each of these, the absorption spectrum shown in Fig. 13-3(A) is
predicted. This is converted to the derivative spectrum in Fig. 13-3(B). This is
an idealized example, and often one finds that the overlapping features generated
by gl and gi make it difficult to obtain their values.
When the system is orthorhombic and g. > q, > g, the powder spectrum
obtained for I = 0 is like that in Fig. 13-4(A). When I = %/2 and the system has
nearly isotropic g-values, but A, > A, > A., the spectrum in Fig. 13-4(B) is
expected. The spectrum for a complex with axial symmetry and I = ', in which
gi > g! and Al > A., is illustrated in Fig. 13-4(C). Other systems become quite
complex, and the possibility for misassignments becomes very large. Only in the
relatively simple cases can the g- and A-values be determined with confidence.
Computer programs are available to simulate powder epr spectra for simple
systems.
Liquid crystal nematic phases can also be used 3 ) to orient a molecule for
epr work. The molecule to be studied, which is the solute, cannot be spherical;
as an example, consider the molecule Co(Meacacen) in Fig. 13-5(A). The liquid
crystal solution of this low spin Co(II) complex is placed in a magnetic field to
orient the liquid crystal molecules (and, in turn, the solute molecules) and is then
cooled. This is schematically illustrated in Fig. 13-5(B). The epr spectrum(4 a) in
zy
H3C / CH 3 (C)
C-o o-C
-H 3 C-C Co C-CH 3 - x
C-N N-C
H3C CH3
(A)
Liquid crystal
Co(meacacen)
(E)
(B)
I I I
H 91 H9 2 3
FIGURE 13-5 The epr spectra (4) of Co(Meacacen) at 770 K. (A) Structural formula;
(B) orientation of the molecule in a frozen oriented liquid crystal; (C) unoriented
frozen solution; (D) frozen liquid crystal oriented as in (B); (E) frozen liquid crystal
reoriented 90 from (B). The phasing in this spectrum is inverted relative to what
one normally employs. [Reprinted with permission from B. M. Hoffman, F. Basolo,
and D. L. Diemente, J. Amer. Chem. Soc. 95, 6497 (1973). Copyright by the
American chemical Society.]
(D) is for the sample oriented relative to the magnetic field as shown in (B), while
in spectrum (E) the sample is rotated 90 around the z-axis (i.e., y parallel to the
field) relative to the magnetic field. Upon rotation, the portion of the spectrum
corresponding to g2 is enhanced, but that for g, is not. One could easily make
the mistake of assuming that this is an axial system with g, assigned to the z-axis
(i.e., g , the axis perpendicular to the plane) and g2 and g3 assigned to gi, where
g, and g, are similar. However, with the molecular coordinate system as defined
in Fig. 13-5(A), g. must be assigned to g3 , g. to g I, and g, to g2 . These assignments
have subsequently been confirmed by a single crystal epr study.(4 b) Difficulties
can arise in this application if care is not taken to demonstrate that the liquid
crystal is not coordinating the complex being studied.
13-2 INTERPRETATION OF THE g-VALUES
Introduction
In contrast to organic free radicals, the g-values of transition metal ions can differ
appreciably from the free electron value of 2.0023. Such deviations provide
considerable information about the electronic structure of the complex. Different
g-values arise because spin-orbit coupling is much greater in many transition
metal ion complexes than in organic free radicals (vide infra). Thus, spin-orbit
effects become essential to an understanding of esr.
The value of g for an unpaired electron in a gaseous atom or ion, for which
Russell-Saunders coupling is applicable, was given earlier by the expression
J(J + 1)+ S(S + 1) - L(L + 1)

2J(J + 1) (13-2)
In condensed phases, first row transition metal ion systems not only do not have
g-values in accord with this expression, but they often deviate from the spin-only
value. In condensed phases, the orbital motion of the electron is strongly perturbed
and the orbital degeneracy, if it existed before application of the chemical
2
A environment, is partly removed or "quenched." If the electron has orbital angular
momentum, the angular momentum tends to be bolstered by being weakly coupled
2B to the spin. There is therefore a competition between the quenching effect of the
ligands-the "crystal field"-and the sustaining effect of the spin-orbit coupling.
Were it not for spin-orbit coupling, we should always observe an isotropic g-value
of 2.0023.
These effects can be illustrated by considering the influence of a crystalline
field on a d' ion as shown for 0 h and D4, (z-axis compression) in Fig. 13-6.
B Equation (13-2) would describe the 2D gaseous ion. The octahedral crystal
field
splits 2D into 2T2 and 2 Eg states. The degenerate T2g state may be further split
by distortion (e.g., Jahn-Teller effects) or by a tetragonal ligand field into E and
Gaseous Oh D 4h B 2 levels. Spin-orbit coupling, on the other hand, tends to preserve a small amount
ion
of orbital angular momentum, so, in the tetragonal complex, the orbital angular
FIGURE 13-6 Splitting of momentum is not completely quenched. One generally refers to this as mixing
the 2 D state b yoh and DO, in of a nearby excited state by spin-orbit coupling. When the amount of spin-orbit
fields (z-axis compression coupling is small compared to the tetragonal distortion (i.e., in the case of large
inD4h. distortions), the mixing can be treated by perturbation theory. In an octahedral
complex, spin-orbit coupling is present in the ground 2T2 , state; in order to

obtain the accuracy needed to understand the epr spectrum, this situation cannot
be adequately treated with perturbation theory. Recall that such a treatment was
employed in Chapter 11 on magnetism.
S = '12 Systems with Orbitally Non-degenerate Ground States

The full Hamiltonian for our system with spin-orbit coupling in a magnetic field
is given by
ft = h(Zeeman) + H(SO) = I (h + g,5) + Li (13-3)
One of the effects of spin-orbit coupling is to modify the simple one-electron

d orbital wave functions. This is described by the A -5 term in the Hamiltonian.
2
For example, the spin wave function for the ground state B2,g of a d' ion in a
tetragonal complex is modified by the spin-orbit interaction ;L -S. From
first-order perturbation theory, the wave function for the Kramers' doublet |>
including spin-orbit effects is given by:
1±t> = N IO) + (1 - N 2) -12 Y n)[_0 I n) (13--4)

MMS E(O) - E(n)
The term I0> is the 2 B 2 g ground state before spin-orbit effects are considered
(i.e., for d' with a tetragonal compression, this one electron is in the d. orbital),
while the summation indicates the contribution made by spin-orbit admixture
of the excited states. In this example, the AE term in the denominator indicates
that the 2 E state will make the largest contribution of all the states that mix in.
We can see from this expression that when there is no orbital angular momentum
mixed into the ground state, I) =| 0>. Evaluation of the matrix elements in
equation (13-4) gives the coefficients necessary to write the appropriate wave
functions. These functions are then used with the Zeeman Hamiltonian in equation
(13-3), i.e.,
' = #^LH + ge#5 -
to set up the 2 x 2 matrix involving |+> and -. Note that we have worked
with the full Hamiltonian in equation (13-3), using the two parts separately. The
;^L -$ term modified the wave function on which we are now operating with the
Zeeman Hamiltonian. The problem is solved by using the raising and lowering
operators. Energies are obtained which are expressed as gpHm,, where g is the
effective g-factor in the direction of the field component H, (i = x, y, z). In this
way, using ^L, $z, and the Zeeman Hamiltonian, we obtain
g = 2.0023 + 2AL ( >- n

E(0) -En
) (13-5)
where I0> is the ground state wave function and I n> is that of one of the n excited
states. Since the ground and excited states corresponding to electronic transitions
in an S = '/2 complex with small spin-orbit coupling are adequately described
by real orbital occupations, the state and orbital designations can be used
interchangeably. The g-value is thus seen to be very dependent upon the mixing
in of the excited state by spin-orbit coupling.
The matrix elements <0 I L |n> are non-zero only when the m, value of |0>
equals the mil value of In>. In a real orbital basis set with
d,2-,2= ( + 2 - 2>)
and
dx, = (+±2) + 2>)
we will have non-zero matrix elements for (d,2 _ I, d,,. (This can be seen
from the triple product F 2 _,2 X F x F,,.) For a d' complex, where we can use
one-electron oribitals instead of states, an evaluation of all the matrix elements
leads to
g = 2.0023 + (13-6)
E(O) - E(n)
where = + /2S for a shell that is less than half-filled. The symbol represents
the one-electron spin-orbit coupling. When the shell is more than half-filled, we
generally think in terms of formation of the positive holes that accompany the
electronic transitions. The sign of the second term, 8 /[E(O) - E(n)], is changed
by changing the denominator to E(n) - E(O). When the shell is less than half-filled,
spin-orbit effects reduce g from 2.0023; but when the shell is more than half-filled,
spin-orbit effects increase g above g. E(0) and E(n) are the energies of the ground
and excited states, respectively.
In a tetragonal complex we have
2, Y (0|Ljln><n|L|0>
g, = 2.0023 + " E(0) - E(n) (13-7)
where g. = g,. These matrix elements are evaluated by using the raising and
lowering operators, so matrix elements are non-zero only when the ground and
excited state m/ values differ by + 1. For a d' tetragonal complex (Fig. 13-6) we
have:
g = 2.0023 + ~ (13-8)
E,, - Exz
The results of the evaluation of matrix elements by this procedure can be
summarized by writing the following general expression for the g-values of S = 12
systems:
g = 2.0023 + E n ) (13-9)
E(0) - E(n)
The values of n are obtained from the so-called magic pentagon shown in Fig.
13-7, which summarizes the results of the evaluation of matrix elements <01 L In>.
m,
z2 0
xz 2yz +1 FIGURE 13-7 The "magic pentagon"

for evaluating n of equation (13-9).
2 2 2 2
x 2 _ Y2( >xy -t2
We will repeat the problem discussed earlier (the d' tetragonal field) with the
use of this pentagon. We first determine that, for an electron in an xy orbital
(i.e., an xy ground state), only electron circulation into the x 2 _ y2 orbital could
give orbital angular momentum along the z-axis. The quantity g. then has
contributions only from xy and x 2 _ y2 , which is seen in Fig. 13-7 to have the
value n =8. Equation (13-9) then becomes:
g = 2.0023 + (13-10)
Ex, - E,2 2
Since gx has non-zero matrix elements only when I 0> and In> differ by Am/ = 1,
we obtain
g = 2.0023 + 2 (13-11)
Ex, - Exz
and
qy = 2.0023 + (13-12)
''Ex, - EZ
These formulas in effect tell us what orbitals permit electron circulations about
the respective axes; that is, x 2 - y2 and xy about z, xy and xz about x, or xy
and yz about y. The "magic pentagon" is easily constructed. The three rows
represent orbitals possessing different m, values; the top row corresponds to
m, = 0, the second row to m/ = i 1, and the third row to m/ = + 2. It should be
emphasized that this whole treatment arises from a perturbation assumption and
is valid only when ng/[E(0) - E(n)] is small compared to the diagonal Zeeman
elements. When second-order perturbation theory is pertinent, a term propor-
tional to (2/AE 2 is added.
Before concluding this section, an application of the use of g-values will be
presented. In Fig. 13-8, the crystal field splittings are indicated for tetragonal
copper(II) complexes with a metal-ligand distance on the fourfold (z)-axis that is
compressed [part (A)] and elongated epart (B)] along this direction. In (A),
the unpaired electron is in dz, with me = 0. There is no other orbital with an
m/ = 0 component, so g = 2.0023. The value of g, = g, is given by
g = 2.0023 + E ''E2 (13-13)

--4- z2 -4- x 2 - Y2
FIGURE 13-8 The d-lIevel splittings -44- x2 - y 2 -4- z2
for a tetragonally compressed
complex (A) and a tetragonally 4- -4- xz, yz -44- xy
elongated complex (B)- -f4- xy -44- -44- xz, yz
(A) (B)
Note that we have switched the order of the energies in the denominator because
the shell is more than half-filled. On the other hand, for a tetragonal elongation
[Fig. 13-8(B)], with the unpaired electron in d,2,2, we obtain
g = 2.0023 + (13-14)
Ex, - E,2 _,2
The value of g, = g, is given by
g, = 2.0023 + 2(13-15)
E,2 - E 2 _,2
Thus, the g-values can be used to distinguish the two structures. Complexes with
a tetragonal compression are very rare, but elongated ones are common.
Copper(II) porphyrin(5') complexes have been reported with g-values of g, = 2.70
and g, = 2.04.
The g-values in some six-coordinate copper(II) complexes exhibit interesting
behavior. As mentioned in Chapter 9, when the single crystal epr spectrum for
[Cu(H 2 0)]SiF,, diluted with the corresponding diamagnetic zinc salt, was
obtained at 90 K, the spectrum was found to consist of one band with partly
resolved hyperfine structure and a nearly isotropic g-value.(6 ) Jahn-Teller distor-
tion is expected, but there are three distortions with the same energy that will
resolve the orbital degeneracy. These are three mutually perpendicular tetragonal
distortions (elongation or compression along the three axes connecting trans
ligands). As a result, three distinguishable epr spectra are expected, one for each
species. Since only one transition was found, it was proposed that the crystal
field resonates among the three distortions (a so-called dynamic Jahn-Teller
distortion). When the temperature is lowered, the spectrum becomes anisotropic
and consists of three sets of lines corresponding to three different copper ion
environments distorted by three differerk. tetragonal distortions.!') The following
parameters were reported for CuSiF, -6H20:(9)
90 K 20K
gi = 2.221 + 0.005 g= 2.46 + 0.01
gi= 2.230 f 0.005 = 2.10 + 0.01
A = 0.0021 + 0.0005 cm-' = 2.10 + 0.01
B = 0.0028 + 0.0005 cm-' A = 0.0110 + 0.0003 cm 1
A< 0.0030 cm-'

Aj
Other mixed copper salts have been found to undergo similar transitions: (Cu,
Mg) 3La 2 (NO 3 )1 2 '24D 2 0 between 33 and 45 K, and (Zn, Cu)(BrO 3)2 6H2 0,
incomplete below 7 K.
A similar behavior (i.e., a resonating crystal field at elevated temperatures)
was detected in,001 the spectra of some tris complexes of copper(II) with
2,2'-dipyridine and 1,10-phenanthroline~") as well as with trisoctamethyl-phos-
phoramide.o 2 >This latter ligand, [(CH 3)2N] 2 P(O)OP(O)[N(CH 3)2] 2 , is a biden-
tate chelate in which the phosphoryl oxygens coordinate. The complex can be
studied without having to dilute it in a diamagnetic host. A complete single
2
crystal epr study and an x-ray diffraction study are reported on this system. ( )
At 90 K, the spectra indicated that at least three different magnetic sites exist,
each of which is described by a g-tensor. Analysis of the spectra indicates that
these sites correspond to distortions along the x-, y-, and z-axes, respectively, in
effect locking in the various extremes in the dynamic Jahn-Teller vibrations
occurring at room temperature. The interesting problem of determining the
chirality of the molecule in a single crystal epr spectral analysis has also been
discussed.!12 )
In single crystal epr studies on transition metal ion systems, it is com-
mon i3-15) to find complexes in which the g- and A-tensors are not diagonal in
the obvious crystal field coordinate system. An axis that is perpendicular to a
reflection plane or lies on a rotation axis must be one of the three principal axes
of the molecule. The g-tensor of the molecule and the A-tensor for any atom
lying on the axis must have principal values along this coordinate. If a molecule
contains only a single axis that meets the above requirements, the other two axes
used as the basis for the crystalfield analysis will not necessarily be the principal
axes of the corresponding tensors of g and A; i.e., selecting these axes may not
lead to a diagonal tensor. For example, bis(diselenocarbamate) copper(II) has
13 4
C 2 h symmetry. , ) The twofold rotation axis is one of the axes that diagonalize
the corresponding g- and A-tensor components, but the other two components
are not diagonal in an axis system corresponding to the axes of the crystal field.
Had the molecule possessed D 2 h symmetry, the three twofold rotation axes of
this point group would have been the principal axes for both the A- and g-tensors.
Thus, epr studies can provide us with information about the symmetry of the
molecule. For a molecule with no symmetry, none of the molecular axes need be
coincident with the axes that diagonalize the g-tensor or A-tensor.* As a matter
of fact, the axis system that diagonalizes A may not diagonalize g. In vitamin
B,2 for example,
15 ) the angle between the principal xy-axis system that leads
to a diagonal A-tensor and those that lead to a diagonal g-tensor is 50'.
Systems in which Spin-Orbit Coupling Is Large

When spin-orbit coupling is large, perturbation theory cannot be used to give
the appropriate wave functions; i.e., equation (13-4) does not apply. An octahedral
d' complex with a 2 T2 g ground state is a case in which spin-orbit coupling is
* For low symmetry molecules, the g- and A-tensors may in fact be asymmetric (i.e., a, # a.).
2
The g -tensor will always be symmetric.
large. When the spin-orbit operator (L 5 operates on the sixfold degenerate real
orbital basis set
1 ,1/2) 2, - 2 1,--
#22
1 = , - - -2, -- -1 2
(the wavefunctions are expressed as |ML, Ms>), the result is a new set of orbitals,
appropriate for d' systems with a large amount of spin-orbit coupling. These are
obtained by evaluating the 6 x 6 determinant containing matrix elements
j ILh - S I j> to obtain the energies. The energies are substituted back into the
secularlike equations to obtain the eigenfunctions pi listed below.
-P (1 ,3)( ,,24k ± 0,6)
92 = (1/ 3)( /22 + 03)
P4= 051
96, = (1/ 3)(41~ - 24016)

The corresponding energies are EI = - /2, E 2 = - /2, E 3 - -/2, E 4 - /2,
E5 = , and E 6 = . We see from this analysis that spin-orbit coupling has
removed the sixfold orbital degeneracy of the T state, giving a twofold set of
levels and a lower-energy fourfold set, corresponding to F, and [, in the 0'
double group.* Next, we need to determine the effect of the magnetic field. Since
systems with 0 symmetry are magnetically isotropic (x, y, and z), it is only
necessary to work out the effect of H.. The Hamiltonian operator, fI (parallel
to z) is #(L + geSz)H_. The resulting energies were derived in Chapter 11 and
are indicated in Fig. 13-9. The splitting in the low-energy set is very small, being
second order in H (i.e., H 2 ). When one solves for g(AE = g#H), the result
g#H = 4#2H/3 or g = 4#H/3 r=" 0 is obtained for the lowest level.t With
appreciable separation of F. and F, (e.g., = 154 cm-' for Ti"), epr will not
be detected unless the F7 state is populated. Solving the expression for g#H for
these levels as above results in a g-value of 2.0. This state will be populated at
room temperature, but the large amount of spin-orbit coupling gives rise to a
short r, and no spectra are observed. Liquid helium temperatures are needed to
lengthen Te, but then only the F states are populated. Thus, the inability to
observe epr spectra on these systems is a direct prediction of crystal field theory.
* The F, and F, states correspond to J = /2 and J = '/2, respectively, for the free ion. These
designations are derived by factoring the product of T2' (orbital part) and E ' (F6 for spin = '/2) in
2
the 0' double group.
tIn some systems, transitions with very small g-values have been detected.
(2 E = + PH, + ( 2H,2
E = { - I3H, + (4f2Hzz/
\3 FIGURE 13-9 The influence
2 T2 of spin-orbit coupling and an
applied magnetic field on a T
state.
(4) r (2) tE
at
E = -(Q/2) - Afl2Hr2/)
Oh Field Spin-Orbit H
Coupling
13-3 HYPERFINE COUPLINGS AND ZERO FIELD

SPLITTINGS
Hyperfine and Zero-Field Effects on the Spectral Appearance

Transition metal systems are rich in information arising from metal hyperfine
coupling and zero-field splitting. Figure 9-14 illustrates the rich cobalt hyperfine
interaction in Co 3 (CO),Se. Before continuing with this discussion, the reader
should review the section on "Anisotropy in the Hyperfine Coupling" in Chapter
9, as well as that on "The epr of Triplet States," if necessary. The spin Hamiltonian
for a single nucleus with spin I and a single effective electronic spin S can be
written to include these extra effects as
fH = #l -g - 5 +S -D -$ + hS -AI - g3NH -I (13-16)
In the previous section, we discussed the #1H g - S term and the complications
introduced by orbital contributions. The next term incorporates the zero-field
effects previously described by the dipolar tensor, D, which has a zero trace. The
last two terms arise when I # 0.
In Chapter 9, we discussed zero-field effects that arose from the dipolar
interaction of the two or more electron spin moments. In transition metal ion
systems, this term is employed to describe any effect that removes the spin
degeneracy, including dipolar interactions and spin-orbital splitting. A low
symmetry crystal field often gives rise to a large zero-field effect.
In an axially symmetric field (i.e., tetragonal or trigonal), the epr spin
Hamiltonian that can be used to fit the observed spectra for effective spin systems
lower than quartet takes the form
lspin = D[$Z2 - S(S + 1) + gj IHZS + g /(H $ + H,$,) +
Aj $,Z, + A1($xIZ + $ ,Z,) + Q'[,2 _ 1(1 + 11 _ Y/3N 0

572 Chapter 13 Electron Paramagnetic Resonance Spectra of Transition Metal [on Complexes
The first term describes the zero-field splitting, the next two terms describe
the effect of the magnetic field on the spin multiplicity remaining after zero-field
splitting, the terms in Al and A, measure the hyperfine splitting parallel and
perpendicular to the main axis, and Q' measures the small changes in the spectrum
produced by the nuclear quadrupole interaction. All of these effects have been
discussed previously (see Chapter 9). The final term takes into account the fact
that the nuclear magnetic moment PN can interact directly with the external field
PNHo = YNHO -I, where y is the nuclear magnetogyric ratio and f#N is the nuclear
Bohr magneton. This is the nuclear Zeeman effect, which gave rise to transitions
in nmr. This interaction can affect the paramagnetic resonance spectrum only
when the unpaired electrons are coupled to the nucleus by nuclear hyperfine or
quadrupole interactions. Even when such coupling occurs, the effect is usually
negligible in comparison with the other terms.
In the case of a distortion of lower symmetry, there are three different
components g., g,, and g., and three different hyperfine interaction constants
A, A,, and A.. Two additional terms need to be included: E($X2 - 5,2) as an
additional zero-field splitting and Q"(I 2 _ 2) as a further quadrupole interac-
tion. The symbols P and P' are often used in place of Q' and Q", respectively.
The importance of the spin Hamiltonian is that it provides a standard
phenomenological way in which the epr spectrum can be described in terms of
a small number of constants. Once values for the constants have been determined
from experiment, calculations relating these parameters to the electronic con-
figurations and the energy states of the ion are often possible.
The splitting of the 6S state of an octahedral manganese(II) complex is
illustrated in Fig. 13-10(A). Here we have the interesting case in which 0 Mn 2
has a 'Asg ground state, which is split by zero-field effects. Spin-orbit coupling
mixes into the ground state excited "T2 states that are split by the crystal field,
and this mixing gives rise to a small zero-field splitting in Mn 2 +. The dipolar
interaction of the electron spins is small in comparison to the higher state mixing
in this complex. The orbital effects are very interesting in this example because
the ground state is 'S, and thus the excited 4 T2 state can only be mixed in by
second order spin-orbit effects. Thus, the zero-field splitting is relatively small,
16
for example, of the order of 0.5 cm-' in certain manganese(II) porphyrins.( a)
As indicated in Fig. 13-10, the zero-field splitting produces three doubly
degenerate spin states, M_ = + 5/2, ± 3/2, ± 1/2, (Kramers' degeneracy). Each of
these is split into two singlets by the applied field, producing six levels. As a
result of this splitting, five transitions (-/22 - -3 ' 22 , 1 -2 '
'/2 -- / 2' /2 -+ '/2) are expected. The spectrum is further split by the nuclear
hyperfine interaction with the manganese nucleus (I = 5/2). This would give rise
to thirty peaks in the spectrum.
In contrast to hyperfine splitting, the term fine splitting is used when an
absorption band is split because of non-degeneracy arising from zero-field
splitting. Components of fine splitting have varying intensities: the intensity is
greatest for the central lines and smallest for the outermost lines. In simple cases,
the separation between lines varies as 3 cos 2 0 - 1, where 0 is again the angle
between the direction of the field and the molecular z-axis.
In Fig. 13-11, the influence of zero-field splitting on an S = 1 system is
indicated for a fixed molecular orientation. (Recall that there is no Kramers'
degeneracy.) In the absence of zero-field effects [Fig. 13-11(A)], the two | Am, =
transitions are degenerate and only one peak would result.
+ FIGURE 13-10 (A) Splitting

of the levels in an octahedral
Mn(lI) spectrum. (B)
Spectrum of a single crystal
of Mn2 + doped into MgV 2 0 6 ,
+ _ _
showing the five allowed
transitions (fine structure),
each split by the magnanese
2 /D nucleus ( = 5/2) (hyperfine
structure). At 300 K,
g, = 2.0042 ± 0.0005,
_ 4S
+ 1 ~ gy = 2.0092 + 0.001, and
2.. 2 - gz = 2.0005 + 0.0005; Az
A= Az= -78+5G; and
D = 218+ 5 G,D = -87 t 5
Gm abdDz = -306 + 20 g.
., 2/ [Modified from H. N. Ng and
C. Calvo, Can. J. Chem., 50,
3619 (1972). Reproduced by
permission of the National
Zero field Applied Nuclear Fine structure Research Council of Canada.]
levels field splitting transitions
50O G
(B)
FIGURE 13-11 The energy

m, =-1,0,+1 level diagram and transitions
for a molecule or ion with
S = 1, (A) in the absence of
and (B) in the presence of
i i i i. zero-field splitting. The
0 Inc.field 0 Inc. field system in (B) is aligned with
strength strength
the z-axis, for the effect is
(A) (B) very anisotropic.
For the splitting shown in Fig. 13-11(B), two transitions would be observed
in the spectrum. A specific example of this type of system is the 3 A g ground
state of nickel(II) in an 0 , field. Spin-orbit coupling mixes in excited states, which
split the 3A,, configuration. Recall that zero-field splitting is very anisotropic,
providing a relaxation mechanism for the electron spin state. Accordingly, epr
spectra of Ni(II) 0 complexes are difficult to detect, and when they are studied,
liquid nitrogen or helium temperatures must generally be employed. At room

temperature, nmr spectra can be measured. In some systems, sharp double
1
quantum transitions (I Am, = 2) can be seen in the epr spectrum. I 6b)
As mentioned in Chapter 9, the zero-field splitting can be so large that the
Am, = + 1 transitions are not observed. For example, in V3 + (d2 ), Am, = + 1
transitions are not detected, but one observes a weak Am, = 2 transition
(-1 -+ +1) split into eight lines by the nuclear spin of "V (1 = 7/2). The
mechanism whereby the Am, = 2 transition becomes allowed is the same as that
discussed for organic triplet states in Chapter 9.
The next topic to be discussed is the definition of the term effective spin, S'.
We have already been using this idea, but now formally define it to describe how
some of the effects that we have discussed are incorporated into the spin
Hamiltonian. When a cubic crystal field leaves an orbitally degenerate ground
state (e.g., a T state), the effect of lower symmetry fields and spin-orbit coupling
will remove this degeneracy as well as the spin degeneracy. In the case of an odd
number of unpaired electrons, Kramers' degeneracy leaves the lowest spin state
doubly degenerate. If the splitting is large, this doublet will be well isolated from
higher-lying doublets. Transitions will then be observed only in the low-lying
doublet, which behaves like a simpler system having S = '/2. We then say that
the system has an effective spin S' of only /2 (S' = 1/2). An example is Co". The
cubic field leaves a 4 F ground state which, as a result of lower symmetry fields
and spin-orbit coupling, gives rise to six doublets. When the lowest doublet is
separated from the next by appreciably more than kT, the effective spin has a
value of '/2 (S' = '/2). instead of 3/2. A spin Hamiltonian can be written in terms
of S' rather than S.
It should be clear that all of the effects discussed above can have a pronounced
influence on the spectral appearance. The qualitative interpretation of the epr
spectra of transition metal ions by inspection is thus not trivial. Proficiency is
obtained by looking at many spectra and drawing analogies to known systems
when dealing with new systems. Practice will be afforded in a later section where
the representative spectra of different d" systems will be considered, and also in
the exercises at the end of the chapter.
Contributions to A
The hyperfine coupling interaction has contributions to it from Fermi contact,
dipolar nuclear spin-electron spin, and nuclear spin-electron orbit mechanisms.
These effects have been discussed in Chapter 9. The reader is referred to equation
(9-25) and the subsequent discussion of it for a treatment of the dipolar
contribution. The trace of the tensor in equation (9-25) is zero, so information
about the dipolarcontributioncan be obtained onlyfrom orderedor partially ordered
systems. As mentioned in equations (9-34) and (9-35), where the traceless tensor
components of A were indicated by T, the contribution to dipolar hyperfine
coupling for an electron in a pz orbital is given by
T (, P, T = - 1P, T,,= - P,
with
Pp -ge#gNN 3 (13-17)
r,
TABLE 13-1. Contributions to the Dipolar Hyperfine

Coupling from Electrons in d orbitals
Orbital TI{Pd) T,,(Pd) TYY(Pd)
d 4 -2/_ 27
d _ -- 4 27 27
d 2 _4 2
2 47
d 2
227 4
d 4/7 27 27,
For an electron in a p, orbital, we have
Tz = - P, T, - P, T ,= (4P,
Similar expressions can be derived for an electron in one of the d orbitals. The
quantitites in Table 13-1 must be multiplied by
Pd = ge3lNIN (13-18)
3
to obtain the dipolar contribution. The signs and magnitudes can be easily
remembered. The orbital is located in the 3 cos 2 0 1 plot of the lines of flux
from the nuclear moment, as shown in Fig. 13-12 for the d., orbital. In this
figure, the z-axis of the molecule is aligned with the z-axis of the field to give T,
as a small positive number. Next rotate the orbital counterclockwise 90 , without
rotating the cone, so that the x-axis of the molecule is aligned parallel to the
field (which is still along the z-axis shown in Fig. 13-12) to give a small positive
Tx. Next, start with Fig. 13-12 and rotate the molecule so that the z-axis of the
molecule is perpendicular to the page and the y-axis of the molecule is parallel
to the field (i.e., the z-axis in Fig. 13-12). Now the lobes of dxl are in the negative
region of the cone and a large negative T, is expected. These rotations correspond
N,
N 550 450
' - FIGURE 13-12 Orientation of the

2
dxz orbital relative to the 3 cos 0 1
cone.
to the three mutually perpendicular orientations of the molecule relative to the

field. With information on the signs and magnitudes of the components of the
hyperfine tensor, one can obtain information about the atomic orbitals in a
complex that make the principal contribution to the molecular orbital containing
the unpaired electron.
The Fermi contact contribution has been discussed in detail in Chapters 9
and 12. Unpaired spin density is felt at the nucleus by direct admixture of the
s orbitals into the m.o. containing the unpaired electron and by spin polarization
of filled inner s orbitals by unpaired electron density in d orbitals. When the
4s orbital of the metal is empty, it can mix into the largely metal d-antibonding
orbital; and if this m.o. contains unpaired spin, the electron partly occupies the
metal 4s orbital.
Spin polarization can have two different results. When an inner Is- or
2s orbital is spin-polarized by a-spin in a 3d orbital, an excess of #-spin results
at the nucleus. When the filled 3s orbital is spin-polarized by an electron with
c-spin in a 3d orbital, a-spin results at the nucleus. The effect of direct
delocalization into the 4s-orbital can be shown by comparing Fermi contact
hyperfine values (AF.C.) determined for various cobalt(II) complexes. It has been
found( 17 a) that AF.C. for sixfold coordination falls in the -30 to -45 G region,
for fivefold coordination in the -5 to -25 G region, and for square planar
fourfold coordination in the 0-G region. One can write an equation summarizing
contributions to AF.C:
AF.C = x(A 4 ) + (1 - x)(AJd (13-19)
where A 4 . is the direct hyperfine interaction of one unpaired electron in a 4s

orbital, 1320 G x g#3; and A 3M is the hyperfine interaction arising from spin-
polarization of filled s orbitals by an unpaired electron in a 3d orbital, -90
G x g#l. Applying this equation to the sixfold, fivefold, and fourfold cobalt(II)
complexes, one obtains x = 3 to 4%, x = 4.5 to 6%, and x = 6.5 % respectively,
for admixture of 4s into the m.o. containing the unpaired electron.
The nuclear spin-electron orbit contribution to the coupling constant is
related to the pseudocontact contribution discussed in Chapter 12. The Hamil-
tonian is
H =-- (iS)(-I) -(-I)(IS)] (13-20)

7
Here the nuclear moment is interacting not only with the spin moment (discussed
earlier), but also with the orbital moment. The nuclear interaction with the spin
moment is a traceless tensor, but the interaction with the orbital moment is not.
Since it is not, a pseudocontact contribution is observed in the nmr spectrum in
solution.
13-4 LIGAND HYPERFINE COUPLINGS
Hyperfine splittings from ligand atoms have contributions from a Fermi contact
term, F.C.; dipolar contributions from the metal, DIP; dipolar contributions from
electron density in p orbitals of the ligands, LDP; and the metal pseudocontact
13-4 Ligand Hyperfine Couplings 577
contribution at the ligand, LPC, which results from the interaction of the orbital
angular momentum of the unpaired electron with the ligand nuclear spin. When
ligand hyperfine structure is resolved, this latter term is generally small compared
to the other contributions. When there is extensive spin-orbit coupling, a large
pseudocontact contribution would be expected, but relaxation effects lead to
difficulty in observing the epr spectrum and consequently the ligand hyperfine
splittings. The All and A, values are given by equations (13-21) and (13-22):
Al = AF.C. + 2 (ADIP + ALDP) (13-21)
A, = AF.C. - (ADIP + ALDP) (13-22)
These equations refer to the parallel and perpendicular components of the ligand
hyperfine tensor. When an investigator measures the ligand hyperfine interaction
(A,) from the fine structure of a metal hyperfine peak (e.g., nitrogen superhyperfine
structure on a cobalt hyperfine peak), difficulty will arise if the two hyperfine
coupling tensors are not diagonal in the same coordinate system. Single-crystal
x-ray diffraction and single-crystal epr studies are necessary for complete
understanding of these systems. If one carries out a solution epr study with the
ligand hyperfine structure resolved, AF.C. can be measured directly. Equations
(13-21) and (13-22) cannot be solved for ADIP and ALDP. However, a reasonable
value for ADIp can be calculated from a knowledge of the structure, by use of
equation (13-23):
ADIP gN##N
N 1
-CM (1323)
a
where a is the metal-ligand distance. Equation (13-23) is derived for circumstances

in which the metal-ligand distance is large compared to the metal nucleus-electron
distance. Where this approximation is poor, other equations have been re-
ported." 8 ) The AF.C. and ALDP values are related to the s orbital contribution c,
and p orbital contribution 2, to the molecular orbital by the following equations:
AF.C. 67r N VI ) 2L2 (1324)
2 1\
ALDP - 2 3N
/ 3 P2 (13-25)
Values of AF.C. and ALDP are reported"" for one electron in s and p orbitals from
which the %s and %p character can be calculated. From these equations, the
ratio %s/%p can be determined to indicate the "hybridization of the ligand" in
the complex. This treatment tacitly assumes that delocalization occurs mainly
by a direct mechanism; i.e., spin polarization contributions are ignored.
The application of these ideas will be illustrated with a few brief examples.
The epr spectrum(2 0) of
O C IC
SxFe S O
indicates Aj(14N)/g# = 15.6 G, A1( 14N)/g# = 12.7 G, and As = 13.7 G. Values

of A reported for a full electron in s and p orbitals are A, /g# = 34.1 G and
A,/g# = 550 G. This leads to a ratio of %p/%s of about 1, which indicates an
sp hybrid and a linear Fe-N-O structure. On the other hand, the ratio of p/s
for the nitrosyl nitrogen (21 of Fe(CN),NO 3 - is 1.6, consistent with a bent
metal-nitrosyl structure (an sp2 hybrid would have had a ratio of 2).
Recently, phosphorus hyperfine couplings were observed(") in the epr spectra
of (C6 H5 )3P and PF 3 adducts of tetraphenylporphyrin cobalt(II). The p/s ratio
was reported as 2.7 in the former case and 0.47 in the latter.
13-5 SURVEY OF THE EPR SPECTRA OF FIRST-ROW

TRANSITION METAL ION COMPLEXES
In this section, we shall briefly survey some of the results of epr studies on various
d" complexes. For a more complete discussion, the reader is referred to references
23 and 24. Before beginning this survey, we should mention that spin-orbitcoupling
provides the dominant mechanism for electron relaxation in these systems. In your
reading, you will find statements like "zero-field splitting causes rapid relaxation"
or "g-value anisotropy leads to a short electron spin lifetime," etc. It is to be
emphasized that these are all manifestations of the effects of spin-orbit coupling
in the molecule. We have previously discussed the relationship of spin-orbit
coupling to these effects. Second and third row transition metal complexes become
increasingly more difficult to study by the epr technique because the spin-orbit
coupling constants are much larger.
d'
The results of studies on ions with the electronic configuration d' can be fit with
the spin Hamiltonian:
EH = #(g.HS + gH,$, + gH S] + AXSXIX + ASI, + AZSZIZ (13-26)
In an octahedral ligand field, the ground state is 2 T2 g and there is considerable

spin-orbit coupling in this ground state. All of the Kramers' doublets from the
2
T2 g term are close in energy and extensively mixed by spin-orbit coupling. This
leads to a short r,. In a tetrahedral ligand field, an 2 E ground state (x 2 __ y 2 , z2)
with no first-order spin-orbit coupling results. In this geometry, admixture of the
13-5 Survey of the EPR Spectra of First-Row Transition Metal Ion Complexes 579
nearby 2 Tg excited states into the ground state by second-order* spin-orbit

coupling provides a mechanism leading to a short spin relaxation time for the
electron and broad absorption lines. The complexes usually must be studied at
temperatures approaching that of liquid helium. The 2 T excited state is split by
spin-orbit coupling. When the ligand field is distorted (e.g., in VO 2 ), the ground
state becomes orbitally singlet and the excited states are well removed. Sharp
epr spectral lines result even at higher temperatures.
Equations have been derived"4 using the d' wave functions and the
appropriate spin Hamiltonian, to relate the g-value in a trigonally distorted
complex to the amount of distortion. The distortion is expressed in terms of 6
(cm-'), the splitting of the 2 T state. A large distortion with = 2000 to 4000
cm-' is found in tris(acetylacetonato) titanium(III). As a result of this splitting,
the electron spin lifetime is increased and one is able to detect the epr spectrum
at room temperature.
d2
Very few examples of epr spectra of these ions in octahedral complexes have been
reported because of the extensive spin-orbit coupling in the 'T, ground state.
Tetrahedral complexes have an 3 A2 ground state, so we would expect longer
relaxation times and more readily observed epr spectra. The spectra of these
systems can be fitted with S = 1 and the spin Hamiltonian
H - g fH S + gjIH,$, + g,3H,$, +
D[S2 2]
2
+ E[, _ 2] + ASZZ + B[SXIX + Si,] (13-27)
VM in an octahedral environment(2 )in an A12 0 3 lattice gave q = 1.92, g, = 1.63,

D = + 7.85, and A = 102. An example of a T complex that has been studied(23 )
is V" in a CdS lattice. At liquid N 2 temperature, a g-value of 1.93 and D of
1 cm-' are observed.
d3
Octahedral d' complexes have an 4 A 2 ground state, which must have a Kramers'
doublet lowest in energy. When the zero-field splitting is small, as shown in Fig.
13-13(A) three transitions can sometimes be detected and the zero-field parameter
can be obtained from the two affected transitions. When the zero-field splitting
is large compared to the spectrometer frequency, only one line will be observed,
as shown in Fig. 13-13(B). In general, the spectra can be fitted to the spin
Hamiltonian as follows:
H = g,3HS, + gIH,S + gf/H,$, + D $S2 ] + E[$ 2

_2]
+ A SZ, + A1, [Z + Si] (13-28)
* In second-order perturbation theory, the E-component of this split state is mixed into the
ground state.
2A 5 6 7
2D (B4 kG
I I I I I I I i
Zero
field i 1 0 1 2 3 4 5 6 7kG
H2
II 2
I I |
(C)
(A) (D)
FIGURE 13-13 (A) Small zero-field and magnetic field splitting of the 4A ground
2
state (field along z) for a d3 case and the resulting spectrum.
(B) Trans-[Cr(C5 H5 N)412]+ in DMF, H20, CH20, CH30H glass31 at 9.3 GHz. D > 0.4
cm - 1, E < 0.01.
(C) Trans-[Cr(C5 HN)4Cl2l+ in DMF, H20, CH2OH glass31 at 9.211 GHz.
(D) Computer simulation 33 ) of (C) with g =gi = 1.99, D =0.164cm-1, E= 0.
[Reprinted with permission from E. Pedersen and H. Toftlund, Inorg. Chem., 13,
1603 (1974). Copyright by the American Chemical Society.1
It is difficult to recognize "typical patterns" for Cr 3 + in some systems.!2" The

spectrum shown in Fig. 13-13(D) was calculated by a computer, using an isotropic
g, D = 0.164 cm-', and E = 0 cm-'. Reference 26 contains many spectra of
tetragonal Cr(III) complexes and a detailed analysis of them.
The d3 system has been very extensively studied, particularly Cr 3 +. In
octahedral complexes, the metal electrons are in t2g orbitals, so ligand hyperfine
couplings are usually small. The g-value for this system is given, according to
crystal field theory, by
8)L
g = 2.0023 - (13-29)
A E(4T2e -*4 A2g)
The ground state, being *A2 g, has no spin-orbit coupling and a small amount is
mixed in via the 4 T2g state. Equation (13-29) differs from those presented earlier
in two ways. The spin-orbit coupling is described by (which can be positive or
negative) and characterizes a state. With more than one unpaired electron, the
energy differences also must be expressed in terms of the energy differences of
the appropriate electronic states. Calculating g for V(H 20) 2 + using AE = 11,800
cm-1 and A = 56 cm- 1 gives a value g = 1.964, which is close to the observed12 1)
value of 1.972. For Cr(H2 O), 3 +, AE 17,400 cm -1, 2 = 91 cm 1, and the
predicted g-value is smaller than the experimental value(28 > of 1.977. In the case
of Mn**, the discrepancy is even larger, with a calculated g = 1.955 and an
experimental value of 1.994. This is in keeping with the fact that the crystal field
approximations are poorer and covalency becomes more important as the charge
of the central ion increases.
d'4
There are very few epr spectra reported for this d-electron configuration. The
ground state in this system, in a weak crystalline 0 , field, which is 5 E, has no
orbital angular momentum, so S is a good quantum number. Zero-field splitting
of the +2, 1, and 0 levels leads to four transitions when the splitting is small,
as shown in Fig. 13-14, and none when the splitting is large. Jahn-Teller distortions
and the accompanying large zero-field splittings that are expected often make it
impossible to see a spectrum.
For low spin d4 complexes, the reader is referred to the d 2 section (recall
the hole formalism).
+2
M, +1
---- 2
5
E
-- 0 0
Oh Zero -i->_
Field Field
-2
FIGURE 13-14 Zero-field and magnetic field splitting of
the 5E ground state (field along z) for a d4 case.
d' Low Spin, S = 12
In a strong ligand field of octahedral symmetry, the ground state is 2 T2 . Spin-orbit

coupling splits this term into three closely spaced Kramers' doublets; however,
epr spectra can be seen only at temperatures close to those of liquid helium
because of the large amount of spin-orbit coupling present. Since there are five
d-electrons in these systems, the situation is analogous to d', except that in this
case we are working with a positive hole. Jahn-Teller forces tend to distort systems
such as MX 6 ", so the g-values contained in equation (13-9) are rarely observed.
The splitting of the free ion doublet state by an 0 , field, a D, distortion, spin-orbit
coupling, and a magnetic field are shown in Fig. 13-15. Since we have non-integral
spin, the double group representations are employed when spin-orbit coupling
is considered, and primed symbols are employed for the representations. If one
FIGURE 13-15 Energy level

diagram for a low spin d5
system in an Oh field and a
D3 field, followed by
spin-orbit coupling and the
magnetic field-induced
splittings. With spin-orbit
effects included, the double
group D3' representations are
employed (primed values).
D3 XL 'S
defines a distortion parameter, 3, as in Fig. 13-15, equations for the g-values can
be derived 2 9 ):
3( + 26)
(13-30)
[ + 23)2 + 8 2]1 2
(23 - 34)
+
(13-31)
[ + 26)2 + 8 22
Note that as 3 -* 0, both gl and g, -+ 0, in accord with the equations in Fig.

13-9. For example, Fe"+ in K 3 Co(CN), exhibits 3 0 ) an epr spectrum with
g-l 0.915 and g, a 2.2. Substituting into equation (13-30) and employing
= -103 cm- ' for the free ion produces a value of 3 & 200 cm- .
Large deviations from octahedral symmetry cause an orbitally singlet state
to lie lowest in energy, well removed from orbitally non-degenerate excited states.
Longer electron relaxation times result, and epr spectra can be observed at higher
temperatures. Examples of such a system are low spin derivatives of ferric
hemoglobin(3 1 ) (Fig. 13-16), which possesses a large tetragonal distortion as a
result of the heme plane. Examples of bases, B, that produce a low spin
environment are N 3 , CN -, and OH -. Experimental g-values for the N 3 - species
are g. = 1.72, g, = 2.22, and g. = 2.80. The large anisotropy in g, and g, is
thought to arise from interaction of a specific iron d orbital with a nitrogen
7norbital from the coordinated histidine group that ties the globin to the heme
unit. EPR spectra 3 2 ) of bipyridyl and phenanthrolene complexes of iron(III),
ruthenium(III), and osmium(III) are analyzed in terms of an energy diagram
globin
similar to Fig. 13-15, but one in which the distortion produces the 2 A state lower
FIGURE 13-16 Schematic in energy than 2 E; i.e., 3 is negative. Spin delocalization onto the ligand was
formula for hemoglobin. studied as a function of the metal in this series.
d5 High Spin
This d-electron configuration has been very thoroughly studied. The high spin
complexes have 'S ground states, and there are no other sextet states. The 4T,
is the closest other term, and second-order spin-orbit coupling effects are needed
to mix in this configuration, so the contributions are small. Thus, the electron
spin lifetime is long and epr spectra are easily detected at room temperature in
all symmetry crystal fields. Furthermore, with an odd number of electrons,
Kramers' degeneracy exists even when there is large zero-field splitting. The
results for high spin complexes are fit by:
=g#H-S+ D 112 3
12
+ E[$ 2
$ ,2] + A-Z
+ a 54 + 5, 4 S+
6116] z
4
_70 180 F L2
18 3 47 5 Z
325
16] (13-32)
The higher-power terms in S arise because the octahedral crystal field operator
couples states with M. values differing by ±4, leading to a more complex basis
set and more non-zero off-diagonal matrix elements.
The splitting of the energy levels and the spectrum expected for an undistorted
octahedral iron(III) complex are shown in Fig. 13-17. Zero field leads to the
degenerate '/2 and '/2 states and a non-degenerate 3/2 state at lower energy by
3a'. The magnetic field further removes the degeneracy giving rise to the transitions
and spectra shown in Fig. 13-17. This spectrum is observed for iron(III) doped
into SrTiO 3 . The iron is coordinated with an undistorted octahedron of oxygen
atoms.
In iron(III) complexes with small tetragonal distortion, D < hv and E = 0.
The energy levels and expected spectrum are illustrated in Fig. 13-18(A). Observed
g-values are very close to 2.00 because of the extremely small amount of spin-orbit
coupling. This fact also allows easy observation of epr spectra at room
temperature. In zero field, the /2 state is lowest with 3/,2 higher by 2D, and the
/2 state above the '/2 state by 5D. With small distortions relative to hv all levels
are populated and five transitions are seen with relative intensities of 5: 8: 9: 8: 5.
If D > hv, the situation shown in Fig. 13-18(B) exists, and only one peak is
seen in the epr corresponding to the transition between |+1/2> and |-'/ 2>.
Even if the higher levels are populated, AM, # 1 for the possible transition and
no spectral bands are observed. The g-values can be calculated, using only /2',
/2> and I /, -'/2> as a basis set and employing the Zeeman Hamiltonian,
H = g '#Hz3z + g #(HS, + H,). When H is parallel to z, we have
|5/2, '/2) 1 /2, - /2)>
='/2,
< /21 1/2ge#Hz 0 (13-33)
<5/2, - /2 0 -'/ 2ge#Hz
Solving equation (13-33) leads to AE - g,#H, and gl = g,

3
2
-
2
H--+
FIGURE 13-17 The splitting of the energy levels (A) and the
spectrum (B) expected for an octahedral iron (111)
complex (H
parallel to a principal axis of the octahedron).
For H parallel to x, after using S* = Sx t iS, we obtain
1'/2, '/2) |1/2, - '/2)>

3
S= (/2, /2| 0 / 2 ge#H (13-34)
3
(K/2, - /2| /2fe Hx 0
Diagonalization of equation (13-34) leads to AE = 3g,#H2 and gi = 3g, & 6.0.

Such a situation is well represented by Fig. 13-16, where B is a weak field
ligand such as F- or H20, which lead to a high spin complex. The zero-field
FIGURE 13-18 Energy

levels and expected
spectrum for a d' ion in a
weak (A) and strong (B)
tetragonal field (H parallel to
the tetragonal axis). [From G.
F. Kokoszka and R. W.
Duerst, Coord. Chem. Rev.,
5, 209 (1970).]
H --
(A)
4D 2
3
2
2D
2 1
2
splitting parameter, D, has been measured for several systems of this type by
examining the far infrared spectrum in a magnetic field. Values ranging from 5
g1l = 2.0
to 20 cm-1 are found for various complexes."') A sample epr spectrum of this
system is that of a frozen aqueous solution of the acid form of ferrimyoglobin g_ = 6.0
shown in Fig. 13-19.
The final case to be considered is one in which a geometric distortion occurs FIGURE 13-19 The X-band
in a complex that removes the axial character. In this case, the rhombic complex epr spectrum of a frozen
is described with the parameters D and E not being zero. If E/D ~ 0 (E ~ 0) aqueous solution (77 K) of
the acid form of
represents axial symmetry, an increase in E/D represents a departure toward
ferrimyoglobin. A similar
rhombic symmetry. The maximum possible distortion is E/D = 0.33, leading to spectrum results for
three equally spaced Kramers' doublets. This Hamiltonian again produces three cytochrome C peroxidase
Kramers' doublets, as shown in Fig. 13-20. Solving this Hamiltonian matrix, (pH = 7.0 at 4.2 K).
(g' = 0.61, 9.7, 0.96)

FIGURE 13-20 Kramers'
doublets in rhombic
symmetry (D and E not equal
to zero) for high spin (g' = 4.29, 4.29, 4.29)
iron(Ill). The three principal
components are listed in
parentheses.
(g' = 9.7, 0.61, 0.86)
using the wave functions that diagonalize it, we find each of the three Kramers'
doublets to be a linear combination of 1 ± 1 ± )2>,and | /2, i /2>
i/2,32>,
Thus, transitions within each Kramers' doublet are allowed; the corresponding
g-values are indicated in Fig. 13-20. The separation between the Kramers'
doublets is large enough that transitions between different ones are not observed,
but at most temperatures all three are significantly populated, and many
3 4
resonances are observed. An example( a) of this situation is Na[Fe(edta)] -4H 20
(where edta is ethylenediamine tetraacetate) diluted in a single crystal of the
analogous Co(III) complex. The spectrum, Fig. 13-21, shows one almost isotropic
transition at g = 4.27 and two very anisotropic ones with principal g-values of
9.64 and 1.10, respectively.
g = 9.48
g = 4.07
g= 1.33
g = 3.94
750 1400 1500
FIGURE 13-21 EPR powder spectrum at 77 K of Fe(Ill) in Co(Ill) DTA. The analysis
of these spectra for D and E provides information on the extent of distortion. Cases
in which the distortions are small produce spectral appearances different from
those shown here. The reader is referred to reference 34 for more details.
6
d
This system has not been studied extensively. Low spin complexes are diamagnetic
4
and high spin 0 complexes are similar to d . High spin iron(II) has a g-value
of 3.49 at 4.2 K and a line width of 500 gauss even at this low temperature.
Spin-orbit coupling in the ground state is very large and there are nearby excited
states that can be mixed in. With a J = 1 ground state, two transitions would
be observed if zero-field effects were small. In a distorted octahedral field, zero-field
effects are large and no epr spectrum is observed. Deoxyhemoglobin is such a
species, and no one has obtained an epr spectrum on this system.
d7
The ground state for an 0 h high spin d7 complex is 'T 1,(F). With extensive
spin-orbit coupling, epr measurements are possible only at low temperatures.
With S = /2 and three orbital components in T, a total of 12 low-lying spin states
result. Low temperatures are needed to observe the epr spectra, because of spin
relaxation problems. Only the low-lying doublet is populated, giving a single
peak, with a g-value of 4.33 from an effective S' = 1/2. Spectra of these systems
have been reviewed.(35 )
In tetrahedral symmetry, cobalt(II) complexes with 'A 2 ground states are
similar to tetrahedral d3 except that the 4 T2 excited state is closer in energy to
the 4 A 2 state in cobalt. With more spin-orbit coupling, broader lines are found
for cobalt(II).
In square planar, square pyramidal, and tetragonal complexes, the ground
state becomes a doublet, S = 1/2 state. With little spin-orbit coupling and no
nearby states with spin-orbit coupling, electron spin lifetimes are long, allowing
observation of the epr at room temperature. The X-band spectra of
Co(DPGB)2 -CH 30H (-C 2 H 2 -) 2 S at 25 C in CH 2 C2/toluene and in a
CH 2Cl2/toluene glass at -180 C is shown(3?) in Fig. 13-22. The DPGB is an
abbreviation for a BF 2 capped bis-diphenylglyoximato ligand system. An isotropic
spectrum results at 25 producing an average g of 2.129. At low temperatures,
an anisotropic spectrum results in the frozen glass. The cobalt hyperfine is nicely
resolved in the parallel region.
The spin Hamiltonian for the low spin d7 system is usually given as:
A' = fg H ,$ + gH, + g H,$] + AXSXT, + A,$,, +Aj$IY (13-35)
In the cases of the five- and six-coordinate tetragonal d7 systems, the unpaired
electron resides in the d,2 orbital. For this electronic configuration, assuming
axial symmetry,
g = 2.0023 (13-36)
6
gi, = 2.0023 - 2 (13-37)
AE(z - xz, yz)
This fact has been used(2s) to study the adduct formation of coordinatively
unsaturated cobalt complexes with varieties of axially coordinated bases, B. Good
overlap between the donor lone pair and d 2 z causes readily observed hyperfine
splitting from bases coordinating via nitrogen or phosphorus atoms. Wayland
has utilized the large gyromagnetic ratio (and hence large hyperfine coupling) of
31
P to obtain hybridization ratios for varieties of PX 3 donors with Co(tetraphenyl-
porphyrin)(2 2) and Co(salen). (36)
As shown in Fig. 13-23, the epr spectra of adducts containing nitrogen donor
ligand orbitals that can mix with the dz2 orbital of cobalt(II) (containing the
unpaired electron) possess nitrogen hyperfine. Three-line hyperfine is seen in the
3150 3350
(g> = 2007
FIGURE 13-22 The X-band

spectra of
Co(DPGB) 2 - CH30H -
(-C 2H2-) 2S. (A) At 25 C in
solution; (B) at -180o C in a
frozen glass.
3200 3500
A(Co)
g11 = 2.007
(B)
complex in Fig. 13-23(A), which contains one acetonitrile and one methanol on
the z-axis and five lines (21 I + 1) are seen in the complex in Fig. 13-23(B)
containing two pyridine ligands on the z-axis. These spectral differences have
been used to probe site isolation in polymers containing basic nitrogen functional
groups by adding a solution of Co(DPGB)2 2CH 3 OH to the polymer.(38) A
three-line nitrogen hyperfine indicates that two nitrogens coordinate and the
complex binds to one site. A five-line pattern indicates that the donor sites are
not isolated. Polymer mobility can permit even a dilute concentration of sites to
behave in a non-isolated manner. The complex Co(DPGB) 2 ' 2CH 3CN can be
used to study basic sites of nuclei with I = 0.
The ground state of the planar, low-spin, four coordinate complexes is more
complex( 3 9,4 0 ) with ground states of d,2 reported(40 ) for symmetrical porphyrins
and d, for unsymmetrical Schiff bases.(40 )
Many five-coordinate cobalt(1I) complexes reversibly bind 02 forming an
end on the 0 2 -bound adduct. The unpaired electron in the mainly d. 2 orbital of
the square pyramidal cobalt(II) complex spin pairs with one of the unpaired
7-
3200 3600
iI
FIGURE 13-23 Frozen glass,

X-band epr of (top)
gl = 2.011
Co(DPGB) 2 - CH30H -CH 3CN;
(bottom) Co (DPGB) 2 -2C 5 H5 N.
A(59Co)
3100 3350
| |
electrons of 02 forming a sigma bond. This bonding description is referred to

as the spin-pairing model.(4 -' The resulting adduct has one unpaired electron in
an adduct molecular orbital that is composed essentially of oxygen a.o.'s. In the
02 adducts, gl > gi as seen in Fig. 13-24, making adduct formation easy to
recognize. This type of spectrum with gl > g- will arise independently of the
charge on 02. The charge is determined by the coefficients in the bonding m.o.,
which contains a pair of electrons. The cobalt hyperfine arises mainly from spin
polarization of the pair of electrons in the Co-0 2 sigma bonding molecular
orbital. Using an approach similar to that described for spin polarization of
organic radicals (Chapter 9), an analysis(41 ) of a series of 02 adducts indicates
that the partial negative charge on the bound 02 varies over a range of -0.1 to
FIGURE 13-24 X-band epr

spectrum of a cobalt(II)
dioxygen adduct in a frozen
50 G
glass.
-0.8. A detailed analysis 4 2 ) of conflicting interpretations of the cobalt hyperfine

in these systems supports the spin-pairing model.
d' High Spin

The ground state of the gaseous ion is 3 F with an orbital singlet state lowest in
an octahedral field. The d-shell is more than half-filled, so spin-orbit coupling
leads to g-values greater than the free electron value. The zero-field splitting
makes it difficult to detect epr spectra except at low temperatures. The g-values
found are usually close to isotropic.
d9
The d9 configuration has been very extensively studied. In an octahedral field,
the ground state is 2 E9. A large Jahn-Teller effect is expected, making observation
of the epr spectrum at room temperature possible. In tetragonal complexes, the
ground state is d _,2 (x- and y-axes pointing at the ligands) and sharp lines
result. Note that the quadrupolar interaction of the copper nucleus (see Chapter
9) can be determined from this experiment. The epr results fit the spin Hamiltonian
f gzH$ + g + H$ + AzSZIz + AX$XZX
+ A~Y+ Q, K 2 1 1(1 + 1)] 9N#04-1 (13-38)
g = 2.223
g = 2.057
9.9 11.5 13.1

Field (kg)
2800 3000 3200 3400 3600 ()
FELD 10
(A)
FIGURE 13-25 "Typical" epr spectra for copper(II)

complexes. (A) A typical solution spectrum for a square
planar Schiff base ligand. [From E. Hasty, T. J. Colburn, and
D. N. Hendrickson, Inorg. Chem., 12, 2414 (1973).] (B) Glass
or powder sample run at Q-band frequencies on an axial
complex. (C) Glass or doped powder spectrum of the
complex in (A) run at X-band frequencies.
FELD (G)
(C)
Some typical copper(II) spectra are shown in Fig. 13-25. In (A), an isotropic
solution spectrum is shown. Both nitrogen and proton ligand hyperfine structures
are seen on the high field peak, but not on the low field peaks. This is attributed
to differences in the relaxation times for the transition, which depend upon the
m-value associated with the transition.(4( 3 The solvent employed influences the
molecular correlation time, which in turn also influences the spectral appear-
ance.(43 )
In (B), an anisotropic spectrum is shown at Q-band frequencies. Such spectra
are observed in glass or powder samples of copper complexes diluted in hosts.
The low field g, and high field gi peaks are well separated. With the higher
microwave energy, the individual peaks are broader so the super-hyperfine
splitting is not detected on the gl peak. In the spectrum in (C), at X-band
frequencies, the gl and g, transitions overlap, but much more ligand hyperfine
structure is detected. As mentioned earlier, the temperature dependence of the
spectra of many copper(II) systems has been interpreted in terms of Jahn-Teller
effects.
13-6 THE EPR OF METAL CLUSTERS
EPR studies of molecules containing two or more metal ions have provided
considerable information on indirect (superexchange) exchange mechanisms. The
spin Hamiltonian is given by(4 4 )
H =S J-S2 (13-39)
where J is a matrix connecting the two spin operators S1 and 32 of metal atoms
I and 2. In the common case of a weak bonding interaction involving isotropic
exchange interactions of the same transition metal ions we can write
It = JS1 - $ 2 (13-40)
As mentioned in Chapter 8, JS1 -$2 groups the energy levels according to
IS1 -S 2 -- Si +S
1 -..S . 2 1 (13-41)
with energies
E(S) = (J/2)1 S(S + 1) - Si(Si + 1) - S 2 (S 2 + 1)| (13-42)
The coefficients indicating the contributions of the spin Hamiltonian parameters

for g and A are calculated(4 5 ) by projecting the spin Hamiltonian parameters of
the individual ions onto the S manifold. For example, g, in the coupled system
as well as A, and A12, the hyperfine coupling of metals 1 and 2 in the coupled
system, is given by(4 6)
ge = cig1 + c2g 2 (13-43)
A,, = CA1 (13-44)
A, = C 2 A2 (13-45)
where gi, g2 , A 1, and A 2 refer to the g- and A-values of the metals 1 and 2 in
an analogous complex containing one paramagnetic center, i.e., an M1 -Zn(II) or
Zn(II)-M 2 complex. The coefficients C1 and C 2 are given by
C1 = S(S + 1) + S(Si + 1) - S 2 (S2 + 1)j/2S(S + 1) (13-46)
C2 = IS(S + 1) + S2(S2 + 1) -- Si(Si + 1)|/ 2S(S + 1) (13-47)

Expressions for the zero-field splitting contributions have been reported.( ) For 40
two S = 1/2 metal ions interacting (S, = 1/2, S2 = /2, S = 1), values of C = 1/2
1
and C 2 = 1/2 result. For an S, = 1/2 atom interacting with an S2 = 1 atom, two
S states result, S = 3/2 and S = 1/2. For the former, the coefficients are C 1 = '/2 and
C2 = 2/3 and for S = 1/2, they are C1 = -1/3 and C 2 = 2/3-
For dimers, in which the two metal ions are identical, e.g., Cu-Cu, the g-values
are identical to the values for a single ion, e.g., Cu-Zn. For practice, one can
calculate the g-values and Acu for a Cu-Mn (S = 2) pair using g2 = 2.32, g = 2.08,
g, = 2.06, and A 2 = 139 for a Cu-Zn pair in a Cu-Zn(py-O)Cl 4 (H 20) 2 com-
plex(4 6,4 7 ) assuming g = 2 for the manganese ion. Values of 1.95, 1.99, 1.94, and
23 should result. Note the greatly reduced copper hyperfine.
The spectrum of a polycrystalline sample of a molecule containing two copper
atoms is illustrated in Fig. 13-26. The main peaks at 2500 and 3700 gauss are
assigned to two gi components. The seven copper hyperfine components [two
Cu (I = 3/2) nuclei] are seen in the low field gl peak. The other high field gi peak
is not seen, for it is out of the spectrometer range. The peak at ~3200 G is
assigned to a Am, = 2 transition. The origin of the doublet will be understood
after discussion of the spin Hamiltonian. The single crystal spectra( 48 ) of molecules
containing two copper(II) atoms with S = 1 are fitted to the spin Hamiltonian
H = /3g H z, + gH $x + gH,$,] + D Sz2 _2]
2
+ E[$ - 32] + A JZ+ A,$,i + AZ$'ZJ (13-48)
N N
/ Cu
CC
N N
1000 2000 3000 4000

Field (G)
(A) (B)
FIGURE 13-26 (A) A schematic representation of a dimeric copper adenine
complex; (B) the polycrystalline epr of this complex.
The results for dimeric copper(II) acetate are(49 ) g. = 2.344, g = 2.053,

g =2.093, D = 0.345 cm-', E = 0.007 cm -', A, = 0.008 cm -1, and A.,
A, < 0.001 cm-'. The influence that the J, D, and E parameters have on the epr
spectrum is illustrated in Fig. 13-27. With JI 260 cm-' in Cu 2(AOc)4 ,
transitions from the S = 0 to S = 1 state are not observed in the epr. The exchange
interaction gives rise to a lower energy S = 0 state, so the intensity of the signals
decreases with decreased temperature. This temperature dependence indicates a
J of -260 cm-', corresponding to a separation of the S = 0 and S = 1 states
of 2J or 520 cm- . In the powder spectrum discussed earlier, split gli and gi
peaks arise from the two AMs = ± 1 transitions averaged over the orientations.
In the relatively rare situation where the exchange parameter J is smaller than
the available microwave energy of the epr experiment, it is possible to see epr
transitions between the S = 1 and S = 0 electronic states of a copper dimer. The
0
first example involved a copper dimer and was reported(' ) for the "outer-sphere"
2E
I -~
EIT
S= 1 -4--
2J
S=O I
FIGURE 13-27 The influence of J interactions and zero-field effects9 on the energy
levels and single-crystal epr spectrum of a molecule containing two d copper atoms.
dimers in [Cu 2 (tren) 2 X2](BPh4 )2 , where X =NCO and NCS- and tren is
2,2',2"-triaminotriethylamine. The "outer-sphere" dimeric association occurs
between two Cu(II) trigonal bipyramids by virtue of hydrogen bonding between
the uncoordinated O(S) end of OCN-(SCN-) nitrogen bonded to one copper
and an N-H proton of the coordinated tren. In the case of X-band epr, J-values
of -0.15 cm-1 to -0.05 cm-1 can be gauged by the observation of singlet-
to-triplet transitions as illustrated in Fig. 13-27.
13-7 DOUBLE RESONANCE AND FOURIER TRANSFORM

EPR TECHNIQUES
Consider an experiment in which the epr transition is broadened by coupling to

an 170 nucleus. If one sweeps the nmr 170 frequencies in a decoupling experiment
and monitors the epr intensity, a plot of this intensity produces the 170 nmr
spectrum. This is referred to as an ENDOR experiment(51 ) (electron-nucleus
double resonance) and combines the inherent sensitivity of the epr experiment
with the resolution of the nmr. One generally performs this experiment at several
different epr frequencies (i.e., g-values) to maximize the information obtained. A
recent report(5 2) involved iron-sulfur cluster interactions with a 17 0-labeled
substrate. A broad epr peak at g = 1.88 is monitored as the 170 frequency was
swept over ~10 MHz. Ten peaks corresponding to two five-line patterns from
non-equivalent 170 ligands (I = '/2) are seen in the ENDOR spectrum even
though no 17 0 hyperfine is observed in the epr.
ELDOR is a double-resonance experiment in which the epr is examined
while another electron spin transition is saturated. This application is similar in
concept to the nmr decoupling experiment. Intensity increases from the nuclear
Overhauser effect can result.
Fourier transform epr provides some of the advantages of FTNMR. The
measurement of electron spin lifetimes is possible. The irradiation time must be
significantly shorter than T, and T2. Using available pulses as short as 10
nanoseconds coupled with a dead time allowance of 50 to 100 nanoseconds permit
the measurement of T-values of 10-6 to 10-7 seconds.(5 1)
A second application of FTEPR is electron spin echo envelope modulation
spectroscopy.! 4 ) One selects a field (i.e., g-value) in the epr spectrum and, with
a sequence of microwave pulses, generates a spin echo. The intensity of the echo
is modulated as a function of the delay between pulses due to the magnetic
interaction of nuclear spins coupled to electron spins. Fourier transform leads
to an ENDOR-like spectrum containing nuclear hyperfine and quadrupole
splittings. In a three-pulse(54c) sequence it is possible to suppress one nuclear
modulation frequency enabling one to eliminate undesired hyperfine couplings
and focus on those of interest. For example, one can see deuterium modulation
while suppressing proton modulation when both nuclei are present in a sample.
REFERENCES 1. See M. Tinkham, "Group Theory and Quantum Mechanics," p. 143, McGraw-Hill,
CITED New York (1964).
2. D. M. S. Bagguley and J. H. E. Griffiths, Nature, 162, 538 (1948).
3. J. P. Fackler and J. A. Smith, J. Amer. Chem. Soc., 92, 5787 (1970); J. P. Fackler, J.
D. Levy, and J. A. Smith, ibid., 94, 2436 (1972).
4. a. B. M. Hoffman, F. Basolo, and D. L. Diemente, J. Amer. Chem. Soc., 95, 6497 (1973).
b. F. Cariati, F. Morazzoni, C. Busetto, D. Del Piero, and A. Zazzetta, J. Chem. Soc.
Dalton, Trans., 556 (1975).
5. P. T. Manohanon and M. T. Rogers in "Electron Spin Resonance of Metal Complexes,"
ed. Teh Fu Yen, Plenum Press, New York, 1969.
6. B. Bleaney and D. J. Ingram, Proc. Phys. Soc., 63, 408 (1950).
7. A. Abragam and M. H. L. Pryce, Proc. Phys. Soc., 63, 409 (1950).
8. B. Bleaney and K. D. Bowers, Proc. Phys. Soc., 65, 667 (1952).
9. B. Bleaney, K. D. Bowers, and R. S. Trenam, Proc. Roy. Soc. (London), A228, 157(1955).
10. H. C. Allen, Jr., G. F. Kokoszka, and R. G. Inskeep, J.Amer. Chem. Soc., 86, 1023 (1964).
11. For other examples, see F. S. Ham in "Electron Paramagnetic Resonance," ed. S.
Geschwind, Plenum Press, New York, 1972.
12. R. C. Koch, M. D. Joesten, and J. H. Venable, Jr., J. Chem. Phys. 59, 6312 (1973).
13. R. Kirmse et al. J. Chem. Phys., 56, 5273 (1972).
14. E. Buluggiu and A. Vera, J. Chem. Phys., 59, 2886 (1973).
15. J. R. Pilbrow and M. E. Winfield, Mol. Phys., 25, 1073 (1973).
16. a. Yonetoni et al., J. Biol. Chem., 245, 2998 (1970).
b. J. W. Orton et al., Proc. Phys. Soc., 78, 554 (1961).
17. a. M. C. R. Symons and J. G. Wilkerson, J. Chem. Soc. (A), 2069 (1971).
b. A. H. Maki, N. Edelstein, A. Davison, and R. H. Holm, J. Amer. Chem. Soc., 86,
4580 (1964).
18. B. A. Goodman and J. B. Raynor, Adv. Inorg. Chem. Radiochem., 13, 135 (1970).
19. P. W. Atkins and M. C. R. Symons, "The Structure of Inorganic Radicals," Elsevier,
New York, 1967.
20. W. V. Sweeney and R. E. Coffman, J. Phys. Chem., 76, 49 (1972).
21. D. A. C. McNeil, J. B. Raynor, and M. C. R. Symons, Proc. Chem. Soc., 364 (1964).
22. B. B. Wayland and M. E. Abd-Elmageed, J. Amer. Chem. Soc., 96, 4809 (1974).
23. R. M. Golding, "Applied Wave Mechanics," Van Nostrand, New York (1969).
24. B. R. McGarvey, "Electron Spin Resonance of Transition Metal Complexes," in
"Transition Metal Chemistry," Vol. 3, pp. 89-201, ed. R. L. Carlin, Marcel Dekker,
New York (1967).
25. S. Foner and W. Low, Phys. Rev., 120, 1585 (1960).
26. E. Pedersen and H. Toftlund, Inorg. Chem., 13, 1603 (1974).
27. R. H. Borcherts and C. Kikuchi, J. Chem. Phys., 40, 2270 (1964).
28. C. K. Jorgensen, Acta Chem. Scand., 8, 1686 (1957).
29. A. Carrington and A. D. McLachlan, "Introduction to Magnetic Resonance,"
Harper & Row, New York, 1967.
30. W. C. Lin, C. A. McDowell, and D. J. Ward, J. Chem. Phys., 49, 2883 (1968).
31. D. J. E. Ingram, "Biological and Biochemical Applications of Electron Spin
Resonance," Plenum Press, New York, 1969,
32. R. E. DeSimone and R. S. Drago, J. Amer. Chem. Soc., 92, 2343 (1970).
33. G. C. Brackett, P. L. Richards, and W. S. Caughey, J. Chem. Phys., 54, 4383 (1971).
34. a. R. Aasa and T. Vanngard, Arkiv Kemi, 24, 331 (1965).
b. R. Aasa, J. Chem. Phys., 52, 3919 (1980).
c. R. D. Dowsing and J. F. Gibson, J. Chem. Phys., 50, 294 (1968).
d. J. R. Pilbrow, J. Mag. Res., 19, 308 (1975).
e. J. Peisach "Frontiers of Biological Science," ed. P. L. Dutton et al., Vol. 2, Academic
f. M. Sato and H. Kon, Chem. Phys., 12, 199 (1976).
35. F. S. Kenedy et al., Biochem. Biophys. Res. Comm., 48, 1533 (1972).
36. B. B. Wayland, M. E. Abd-Elmageed, and L. F. Mehne, submitted.
37. B. Tovrog, D. J. Kitko, and R. S. Drago, J. Amer. Chem. Soc., 98, 5144 (1976).
38. R. S. Drago and J. H. Gaul, Inog. Chem., 18, 2019 (1979).
39. F. Urbach, J. Amer. Chem. Soc., 96, 5063 (1974).
40. a. B. R. McGarvey, Can. J. Chem., 53, 2498 (1975).

b. V. Malatesta and B. R. McGarvey Can. J. Chem., 53, 3791 (1975).
c. A. Reuveni, V. Malatesta, and B. R. McGarvey Can. J. Chem., 55, 70 (1977).
41. R. S. Drago and B. B. Corden, Accts. Chem. Res., 13, 353 (1980) and references therein.
42. T. D. Smith and J. R. Pilbrow, Coord. Chem. Rev., 39, 295 (1981).
43. W. B. Lewis and L. 0. Morgan, "Transition Metal Chemistry," Vol. 4, p. 33, ed. R.
L. Carlin, Dekker, New York, 1968.
44. T. Maryia in "Magnetism" eds. G. T. Rado and H. Suhl, Vol. 1, Academic Press,
New York, 1963.
45. J. Scaringe, D. J. Hodgson, and W. E. Hatfield, Mol. Phys., 35, 701 (1978).
46. a. D. Gatteschi "The Coordination Chemistry of Metalloenzymes," eds. I. Bertini, R.
S. Drago, and C. Luchinat, D. Reidel, Holland, 1983.
b. L. Banci, A. Bencini, and D. Gatteschi, Inorg. Chem., 20, 2734 (1981).
47. J. A. Paulson, D. A. Knost, G. L. McPherson, R. D. Rogers, J. L. Atwood, Inorg.
Chem., 19, 2519 (1980).
48. G. F. Kokoszka and R. W. Duerst, Coord. Chem. Rev., 5, 209 (1970).
49. B. Bleaney and K. D. Bowers, Proc. Roy. Soc. (London), A214, 451 (1952).
50. D. M. Duggan and D. N. Hendrickson, Inorg. Chem., 13, 2929 (1974).
51. J. S. Hyde, R. C. Sneed, and G. H. Rist, J. Chem. Phys., 51, 1404 (1969) and references
therein.
52. J. Telser, M. H. Emptage, H. Merkle, M. C. Kennedy, H. Beinert, and B. M. Hoffman,
J. Biol. Chem., 261, 4840 (1986).
53. a. R. Orbach and H. J. Stapleton, "Electron Spin-Lattice Relaxation," in "Electron
Paramagnetic Resonance" ed. by S. Geschwind, Plenum Press, New York 1972.
b. H. J. Stapleton, Mag. Reson. Rev. 1, 65 (1972).
54. a. L. Kevan in "Time Domain Electron Spin Resonance," eds. L. Kevan and R. N.
Schwartz, Wiley-Interscience, New York, 1979.
b. L. Kevan, J. Phys. Chem., 85, 1628 (1981).
c. L. Kevan, Accts Chem. Res., 20, 1 (1987).
d. W. B. Mims and J. Peisach, J. Biol. Mag. Reson., Vol. 3, eds. J. Berliner and J. Ruben,
ADDITIONAL J. E. Wertz and J. R. Bolton, "Electron Spin Resonance," McGraw-Hill, New York, 1972.
A. Abragam and B. Bleaney, "EPR of Transition Ions," Clarendon Press, Oxford, England,
REFERENCES 1970.
J. W. Orton, "Electron Paramagnetic Resonance," Gordon and Breach, New York, 1968.
R. M. Golding, "Applied Wave Mechanics," Van Nostrand, New York, 1969.
J. S. Griffith, "The Theory of Transition-Metal Ions," Cambridge University Press, New
York, 1961.
A. Carrington and A. D. McLachlan, "Introduction to Magnetic Resonance," Harper &
Row, New York, 1967.
C. P. Poole, "Electron Spin Resonance," Interscience, New York, 1967.
B. R. McGarvey, "Electron Spin Resonance of Transition Metal Complexes," in "Transition
Metal Chemistry," Vol. 3, pp. 89-201, ed. R. L. Carlin, Dekker, New York, 1967.
G. Kokoszka and G. Gordon, "Technique of Inorganic Chemistry," Vol 7, pp. 151-271, ed.
H. B. Jonassen and A. Weissberger, Interscience, New York, 1968.
B. A. Goodman and J. B. Raynor, Adv. Inorg. Chem. Radiochem., 13, 135 (1970).
R. Bersohn, "Determination of Organic Structures by Physical Methods," Vol. 2, eds. F.
C. Nachod and W. D. Phillips, Academic Press, New York, 1961.
A. Carrington, Quart. Rev., 17, 67 (1963).
A. Carrington and H. C. Longuet-Higgins, Quart. Rev., 14, 427 (1960).
Exercises 597
1. Below is a typical Cu" solution spectrum. It can be characterized by g =(1/3)g EXERCISES

+ 2/ 3g, and A = /3 All + 2 /3 A,-
H(gauss)
a. Determine g and A.
b. This is an X-band spectrum (9.4 x 10' Hz). Why should A be expressed in Hertz?
2. Readers often find it difficult to understand why the epr spectrum of a powder or
frozen solution should yield (in an axially symmetric compound, for example) g , g-,
A , and A_. "After all," the argument goes, "you are looking at a superposition of
all possible orientations. At intermediate orientations you have an average g and A,
e.g., g2 = g cos 2 0 + g12 sin 2 0. Thus, the spectrum should be very broad with few
features,." This argument is correct as far as it goes, but it ignores the fact that we
are looking at a first derivative. For example, suppose that a compound has g = 2.0
and gi = 2.1. Recalling that there are many more perpendicular directions than there
are parallel ones, the absorption spectrum should appear in the ideal case as in (A):
2.0 2.1
Because of line broadening, the real absorption curve resembles that in (B). The first
derivative will then be
L 2.1
2.0
The argument is further complicated, but essentially unchanged, by including hyperfine

coupling.
Below is the powder spectrum of copper diethyldithiophosphinate (for 63 Cu,
I =/2). Explain as many features as possible. Hint: AI(Cu) ' 5A (Cu). What other
nuclei might give hyperfine splitting?
3. The solution spectrum below is typical for a vanadyl complex (for V, I= ?/2)-
Explain the splitting pattern.
Exercises 599
4. Below are the frozen solution (top) and room temperature (bottom) spectra of a
vanadyl complex of d,l-tartrate (only one species is present). What can you deduce
about the structure of this complex? (Tartaric acid has the formula
H H
I |
HOOC-C- C-COOH.)
O |
OH OH
250 G
X 25
H
1000 2000 3000
100 G
H-+
, integrated
5. Diphenylpicrylhydrazyl (DPPH) is a common reference in epr spectroscopy. Its

structure is
structure is
a N -N
2
OrN
In powder or concentrated xylene solution, it gives a sharp one-line pattern shown

at top left. In a 10- 3 -M xylene solution, it gives the pattern shown at the left. How
does this splitting pattern arise? Why is it present only in dilute solution?
6. Copper(II) tetraphenylporphyrin [CuTPP; see part (D) below.] has a square planar
D 4h structure with the copper lying in the plane of the four equivalent porphyrin
nitrogens. EPR spectra of a sample of CuTPP doped into the free ligand are shown
63 Cu and 31%
H-- below.
65
This sample is 100% 6 3Cu. Natural abundance copper is 69%
Cu, and I = '2 for both isotopes.
67.6 G
a. Spectrum (A) is a powder spectrum obtained at a frequency of 9.4 GHz (X-band),

while spectrum (B) was obtained at 35 GHz (Q-band). Why are ga and g, well
separated in (B) compared with those in (A)?
b. Explain the source of the number of peaks observed in the gj region in spectrum
(B). Also explain the splitting pattern found for each peak [see insert (C)].
Exercises 601
c. What is the crystal field splitting diagram for a square planar complex? How is
observation of the splittings in insert (C) consistent with the orbital occupancy
predicted for a Cu(II) complex by your crystal field diagram?
d. Qualitatively, if this copper compound did not have a threefold or higher axis of
symmetry, what would the powder spectrum look like?
e. If natural abundance copper were used, what effect would this have on the
appearance of insert (C)?
Note that insert (C) is obtained by increasing the gain for the set of low-field lines.
7. Below is a spectrum of cobalt phthalocyanine (for Co, I 2), a D4, Co(II) porphyrin
similar to the copper porphyrin in problem 6.
8
c 7
06
545 magnetic dilution 1 :1000
3 -4
1 2
~0
g =2.041
-+freeradica
1' 2' 3' 4' ;5' 6' 7' 8'
1400 1700 2009 2300 2600 2900 3200 3500 3800 4100 H (gauss)
a. What are the values of g , g , A , and Ai? What is the origin of the observed
hyperfine structure?
b. Dissolving cobalt phthalocyanine in 4-methylpyridine produces a 1:1 adduct. The

frozen solution spectrum at 77 K is shown below. Utilizing the D4, splitting
diagram for low spin Co(II), state what orbital the unpaired electron occupies.
Why is each of the eight upfield components split into three lines?
3 1 _ 1L 3 _5 7 -M
2 - 2 2 2 2 2 M
c. Why do you see super-hyperfine interaction from the pyr-idine nitrogen but not from
the four phthalocyanine donor nitrogen atoms?
8. Predict the number of spectral lines for
a. Co(H2 O), 2 +
b. Cr(H2 O) 63+
Indicate how zero-field splitting and Kramers' degeneracy applies in these examples.
9. The following epr data on a series of high-spin octahedral metal hexafluoride complexes
are taken from Proc. Roy. Soc. (London), 236, 535 (1956):
A metal AF
Complex g x 104 cm-l x 10 cm-1 /met Temp.
MnF 6 4 - g - g, - g, = 2.00 96 17 55
Mn = /2 300 K
CoF 6 4 ~ g = 2.6 A = 43 A = 20 59Co = 7/2 20 K
g, = 6.05 A = 217 A, = 32
g, = 4.1 A = 67 A = 21
CrF 6 3 - g = 2.00 A = 16.2 A = 3 53 Cr
= 3/2 300 K
g = 1.98 A =16.9 A = 1
a. Why do CrF0 3- and MnF6,* give sharp epr at room temperature while CoF 6 4 -
does so only at 20 K? What effect or effects cause these differences in ability
to observe the epr? Which of these complexes would be best for a room temperature
nmr study?
b. Why are the CrF 6 3- and MnF 6 *- g-value fairly isotropic and close to 2.0 while the
CoF 6 '- values are anisotropic and deviate from 2.0?
c. Why do CoF 6 *- and MnF 6 4 have larger AF values than CrF6 3-?
10. The November, 1973, issue of InorganicChemistry reports the liquid and solid solution
epr spectra of some 10-3 M vanadyl dithiophosphinate complexes:
0
RON S" S 'OR
7
RO I-'S S ' OR
?4$44v- W.4J~IVJ~
liquid solid
a. Explain the liquid spectrum, for which R = CH 3 . (The lower one is a computer
simulation.) What parameters are needed to characterize it? V has I = 7/2, P has
I = J/, and H has I = '/2-
b. Explain the solid solution spectrum, for which R = phenyl. What parameters are
needed to characterize it?
Exercises 603
c. What should be the mathematical relationship between the parameters in part a and
those in part b?
11. The spectrum below is that of an axially symmetric Cu 2 + complex in a frozen solution.
It is a d9 system. Assume 100% abundance for 63Cu (I =2). The following constants
will be needed: # = 9.27 x 10-21 erg/gauss; h = 6.67 x 10-27 erg sec; v = 9.12 x 10'
Hz for the spectrometer on which the spectrum was obtained.
2720 2810 2900 gauss
a. How many parameters are required to explain this spectrum? What features of
the spectrum suggest these parameters?
b. Suggest why the high-field half of the spectrum (_Lregion) is more intense.
c. What is the absolute value of g,? Justify the terms in the equation you use.
12. Assume that the square pyramidal complex Cu(hfac) 2 P(CH,) 3 (for 3 1 P, I 1/2; for
63
Cu, I =/ can exist as either of the following isomers [Inorg. Chem., 13, 2517
(1974)]:
P(C6 H5 )3
C (II) /Cu(II)
O ) P(C6H5)3
axial basal
What might be learned to aid in distinguishing the isomers by looking at
a. the electronic spectrum?
b. the infrared and Raman spectra (answer in a general way; i.e., don't work out the
total representation)?
c. the epr spectrum?
d. an 17 0-labeled epr spectrum (for 170, I = 5/2)?
e. the nmr spectrum?

Nuclear Quadrupole
14 Resonance
Spectroscopy, NQR*
14-1 INTRODUCTION
When a nucleus with an electric quadrupole moment (nuclear spin I >; 1; see the
second paragraph of Chapter 7 and Fig. 7-1) is surrounded by a non-spherical
electron distribution, the quadrupolar nucleus will interact with the electricfield
gradient from the asymmetric electron cloud to an extent that is different for the
various possible orientations of the elliptical quadrupolar nucleus. Since the
quadrupole moment arises from an unsymmetric distribution of electric charge
in the nucleus, it is an electric quadrupole moment rather than a magnetic
moment that concerns us. The allowed nuclear orientations are quantized with
21 + 1 orientations, described by the nuclear magnetic quantum number m,
where m has values +I to -I differing by integer values. The quadrupole energy
level that is lowest in energy corresponds to the orientation in which the greatest
amount of positive nuclear charge is closest to the greatest density of negative
charge in the electron environment. The energy differences of the various nuclear
orientations are not very great, and at room temperature a distribution of
orientations is populated in a group of molecules.
If the electron environment around the quadrupolar nucleus is spherical (as
in free Cl~), all nuclear orientations are equivalent and the corresponding
quadrupole energy states are degenerate. If the nucleus is spherical (I = 0 or '/2),
there are no quadrupole energy states even with a non-spherical electron cloud.
In nqr spectroscopy, we study the energy differences of the non-degenerate
nuclear orientations. These energy differences generally correspond to the radio
frequency region of the spectrum, i.e., -0.1 to 700 MHz.
It is helpful to consider the interaction of charges, dipoles, and quadrupoles
with negative electron density in order to define some terms important for nqr
(and M6ssbauer spectroscopy, Chapter 15). In Fig. 14-1(A), we illustrate the
interaction of a positive charge on the z-axis with negative electron density.
The energy is given by -e 2/r or -- eV, where V = -e/r is the electron potential
felt by the positive charge located at the point r. In Fig. 14-1(B), we represent
a dipole moment in the field of electronic charge. Now the energy associated
* Several of the principles covered in Sections 14-1 to 14-4 are common to both nqr and
604 Mbssbauer spectroscopy, so these sections should be read before Chapter 15.
+ 'FIGURE 14-1 The

+ interaction of (A) a positive
charge, (B) a dipole, and (C)
r a quadrupole with the
z-component of the electric
field arising from a unit
0.
negative charge
z z z
(A) (B) (Cl
with the orientation of the dipole depends upon how the potential energy
changes over the dipole. Thus, we are interested in how the electrostatic
potential changes over the dipole or OV/az. This is referred to as the z-
component of the electric field, E.. In Fig. 14-1(C), we illustrate the interaction
of a quadrupolar distribution with the electric field. Now we have the electric
field from the electron in effect interacting with two dipoles whose configuration
relative to each other is fixed, i.e., a quadrupole. The energy will depend upon
the rate of change (or gradient) of the electric field over the quadrupole. Thus,
we are concerned with the "change in the change" of the potential from the
2
electron, or the second derivative of V with respect to z, that is, 8 V/Bz . This
2
quantity, which is also the change in the electric field component, BE,/Bz, is
called the electric field gradient.
In our molecule we have a nucleus imbedded in a charge cloud of electron
density. The electric field gradient is defined in terms of a time-averaged electric
potential from an electron. Furthermore, the electron field gradient is described
by a 3 x 3 tensor V, which is symmetric and has a zero trace. The nuclear
quadrupole moment is also described by a 3 x 3 tensor Q. The nuclear quadru-
pole coupling energy EQ is given by
e
EQ =- 6i;Vij
where Qij is the nuclear quadrupole moment tensor and Vi1 is the electrostatic
field gradient tensor arising from the electron distribution. The product will
depend upon the mutual orientation of the two axis systems. For Q, it is
convenient to select an axis system that coincides with that of the spin system.
When this is done, the cylindrical symmetry of the nucleus permits definition
of the tensor in terms of one parameter, the nuclear quadrupole moment Q.
The sign of this quantity must be known to obtain the sign of the electric field
gradient. An axis system can also be chosen in which the electrostatic field
gradient tensor is diagonal. This is called the principalaxis system of the field
606 Chapter 14 Nuclear Quadrupole Resonance Spectroscopy, NQR
gradienttensor, and the only non-zero elements are the diagonal elements whose
magnitudes produce a traceless tensor:
82 7 + g2 y + g2 v =0
+XOy2 OZ2
When comparing a given atom in different molecules, it is necessary to know

the orientation of the field gradient principal axis system in the molecular
framework axis system. Three Eulerian angles (2, f#, and y) are required.
The asymmetry of a molecule and the direction of the z-axis of the field
gradient, qzz, relative to the crystal axes can be investigated by studying the
nqr spectrum of a single crystal in a magnetic field. The Zeeman splitting is a
function of orientation, and detailed analysis of the spectra for different
orientations enables one to determine the direction of the z-axis of the field
gradient, q... This axis can be compared to the crystal axes.
When the principal axes of the coordinate system of the molecule are
principal axes of the electric field gradient tensor, the potential energy EQ for
interaction of the quadrupole moment with the electric field gradient at the
nucleus is given by
EQ = (VQ + V,Q,, + VzzQZZ) (14-1)
We define the electric field gradient V., as eqz,, where e is the electron charge
(4.8 x 10 -" esu). Since the trace of the electric field gradient tensor is zero, we
need define only one more quantity to specify the field gradient, and this is
done in equation (14-2):
V.=" "
(14-2)
The quantity q is called the asymmetry parameter.The quantities V.,, V,,, and
Vz2 are often written as eq,,, eqy, and eqz2. By convention, Iq..l > Iq I > Iq,,I,
so q ranges from 0 to 1 as a result. With a zero trace, the field gradient is
completely defined by eq and q.
Substituting these definitions of q. and t7 into equation (14-1) and using
the fact that qz2 + q , + q,, = 0, we obtain
EQ = 6 q - (17+ 1)Q,, + Qzz

_ 2)Q. (14-3)
When axial symmetry pertains, q equals zero and EQ becomes equal to

('1 4 )e2 qQz2. (Note that Q2Z + QXX + Q,, = 0.)
The classical considerations given above are readily expressed by a quan-
tum mechanical operator:
HQ
-(2) Y ~ 1-4)
H _= qiQi (14-4
(6 j
where the summation is over the components of the nuclear quadrupole
moments Qi and the electric field gradients qi. In the principal axis system
with q defined as above, the most common form of the Hamiltonian is given by:
e2Qq 1
HQ = 31(22 -- 1(I + 1) -2(f+ + f2)(14-5)
The product e2 Qq or e2 Qq/h (often written as eQq,2 or eQq,,/h) is called the

quadrupole coupling constant. The operator HQ operates on the nuclear wave
functions. When q = 0, the terms involving the raising and lowering operators
drop out. We shall not be concerned with the explicit evaluation of matrix
elements; the interested reader can consult references 1-3. Suffice it to say that
a series of secular equations can be written and solved to give the energies of
the nuclear spin states in the electric field gradient resulting from the distribu-
tion of the electron density of the molecule.
14-2 ENERGIES OF THE QUADRUPOLE TRANSITIONS
In an axially symmetric field (q = 0), the energies of the various quadrupolar

nuclear states are summarized by the following equation:
e2Qq[3m 2 - I(I + 1)](14-6)

41(21 - 1)
where I is the nuclear spin quantum number, and m is the nuclear magnetic
quantum number. For a nuclear spin of I = '/2, m can have values of /2, '/2,
- /2, -3/2. For m = 3/2, substitution into equation (14-6) produces the
result
E3/2 = +e 2Qq/4. Since m is squared, the value for m = _ 3/2 will be identical to
that for m= +3 /2, and a doubly degenerate set of quadrupole energy states
results. Similarly, the state from m = '1/2 will be double degenerate. The
transition energy, zE, indicated by the arrow in Fig. 14-2, corresponds to
e 2Qq 4 - (_ 2 Qq 4) 2Qq/2. Thus, for a nucleus with a spin I = '/2 in an
axially symmetric field, a single transition is expected, and the quantity e 2 Qq
expressed in energy units can be calculated directly from the frequency of
absorption e 2 Qq = 2AE = 2hv. The quantity e 2 Qq is often expressed as a fre-
E- + e 2Qq/4
E 23
E FIGURE 14-2 Quadrupole
E energy levels in a spherical
R AE field (A) and an axially
G symmetric field (B).
Y
E= -e 2Qq/4
608 Chapter 14 Nuclear Quadrupole Resonance Spectroscopy, NOR
quency in MHz, although strictly speaking this should be e2 Qq/h. For the above
case, e2 Qq would be twice the frequency of the nqr transition.
The number of transitions and the relationship of the frequency of the
transition to e2 Qq can be calculated in a similar manner for other nuclei with
different I values in axially symmetric fields by using equation (14-6). For I = ?2',
four energy levels (Et ,' E +3/2, E + 2, and E + 7/2) and three transitions result.
The selection rule for these transitions is Am = i1, so the observed transitions
are Ei 1/2- E 3 /2 , E+ 3 2 -* E+ 2 , and E+ 5 2 -+ E, 7 2. (Recall that all levels are
populated under ordinary conditions). Substitution of I and m into equation
(14-6) produces the result that the energy of the E+ 3/2 - E+ 5/2 transition is
twice that of the E+ 1 /2 - E± 3/2 transition. The energies of these levels and the
influence of the asymmetry parameter rqon these energies are illustrated in Fig.
14-3. In measured spectra, deviations from the frequencies predicted when r = 0
are attributed to deviations from axial symmetry in the sample, and, as will be
seen shortly, this deviation can be used as a measure of asymmetry.
m = ±7/2
> m =±5/2
FIGURE 14-3 Nuclear quad rupole T'
energy level diagram for / - 7/2.
/2-~ ~ M =L 3/2 V
m =±1/2
0 0.2 0.4 0.6 0.8 1.0
In a nuclear quadrupole resonance (nqr)* experiment, radiation in the radio

frequency region is employed to effect transitions among the various orientations
of a quadrupolar nucleus in a non-spherical field. The experiment is generally
carried out on a powder. Different orientations of the small crystals relative to
the r.f. frequency direction affect only the intensities of the transitions but not
their energies. Structural information about a compound can be obtained by
considering how different structural and electronic effects influence the asym-
metry of the electron environment. One set of resonances is expected for each
chemically or crystallographically inequivalent quadrupolar nucleus. Crystallo-
graphic splittings are often small compared to splittings from chemical non-
equivalence.
Two types of oscillators have been commonly used in nqr, the super-
regenerative and the marginal oscillator. The superregenerative oscillator is most
common because it allows broad band scanning in searching for resonances and
is not complicated to operate. It has the disadvantage of producing a multiplet
of lines for each resonance, as shown in Fig. 14-4(A), because of its particular
* For reviews of nqr, see the General References.

(A) (B)
FIGURE 14-4 (A) Multiplet of peaks for a single resonance from a

superregenerative oscillator; (B) singe peak from a marginal oscillator for the
same resonance as in (A).
operational characteristics. The true resonance frequency is the center line of the
multiplet. The marginal oscillator gives a single peak for each absorption, but it
requires constant adjustment and is tedious to operate. Details regarding the
instrumentation have been published.(')
The "Cl nqr spectrum of CI3 BOPCl 3 is shown in Fig. 14-5. Resonances
for three non-equivalent chlorines (Cl 3POBCl 2 Cl-) are found. Two of the
resonance centers are indicated by x marks on the spectrum at 30.880 MHz
and 31,280 MHz. The center of the third resonance around 30.950 MHz is
difficult to determine accurately because of overlap with the resonance at
31.280 MHz.
In addition to this direct measurement of the quadrupolar energy level
difference by absorption of radio frequency radiation, the same information
may also be obtained from the fine structure in the pure rotation (microwave)
I I I I i I i I
6 31.500 4 3 2 1 31.000 9 8 7 6 30.500
MHz MHz MHz
FIGURE 14-5 The 35CI nqr spectrum of Cl3BOPCl 3 at 77 K with 25-kHz markers.
spectrum of a gas. The different nuclear orientations give slightly different

moments of inertia, resulting in fine structure in the microwave spectrum. The
direct measurement by absorption of radio frequency radiation must be carried
out on a solid. In a liquid or even in some solids (especially near the melting
point), collisions and vibrations modulate the electric field gradient to such an
extent that the lifetime of a quadrupole state becomes very short. This leads to
uncertainty broadening, and the line is often not detected.
The energy difference between the various levels and, hence, the frequency
of the transition will depend upon both the field gradient, q, produced by the
valence electrons and the quadrupole moment of the nucleus. The quadrupole
moment, eQ, is a measure of the deviation of the electric charge distribution of
the nucleus from spherical symmetry. For a given isotope, eQ is a constant, and
56
values for many isotopes can be obtained from several sources.( ' ') They can
be measured with atomic beam experiments. The units of eQ are charge times
distance squared, but it is common to express the moment simply as Q in units
of cm2 . For example, "Cl with a nuclear spin I = '/2 has a quadrupole moment
Q of -0.08 x 10 4 cm 2 , the negative sign indicating that the charge distribu-
tion is flattened relative to the spin axis (see Fig. 7-1).
The second factor determining the extent of splitting of the quadrupole
energy levels is the field gradient, q, at the nucleus produced by the electron
distribution in the molecule. The splitting of a quadrupole level will be related
to the product e2 Qq. For a molecule with axial symmetry, q often lies along
the highest-fold symmetry axis, and when eQ is known, one can obtain the
value of q. In a non-symmetric environment, the energies of the various
quadrupole levels are no longer given by equation (14-6), because the full
Hamiltonian in equation (14-5) must be used. For I = 3/2, the following
equations can be derived(') for the energies of the two states:
2 3e 2 Qq 1 + q 2 /3 (14-7)
41(21 - 1)
-3e 2
Qq 1 + /3 (1
41(21- 1)
From the difference of equations (14-7) and (14-8) and E = hv, it is seen that
one transition with frequency v = (e2 Qq/2h) 1 + q2 /3 is expected for I = 3/2-
Since there are two unknowns, ? and q, the value for e 2 Qq cannot be obtained
from a measured frequency. As will be seen shortly, this problem can be solved
with results from nqr experiments on a sample in a weak magnetic field.
The equations for nuclei with I values other than 3/2 have been reported,
and in many instances where more than one line is observed in the spectrum,
both the asymmetry parameter, i, and e 2 Qq can be obtained from the spectrum.
For I = 1, the equations are:
2
E, =-2e Qq (14-9)
41(21 - 1)
14-3 Effect of a Magnetic Field on the Spectra 611
and
= e2 Qq(1 + q) (14-10)
41(21 - 1)
The corresponding energy levels are indicated in Fig. 14-6(B), where K stands
for e 2 Qq/(41(2I - 1)). The perturbation of these levels by an applied magnetic
field is indicated in Fig 14-6(C). This effect will be discussed in the next section.
As can be seen from Fig. 14-6(B), there are three transitions for I = 1 and q # 0
(labeled vi, v , and v,), so the two unknowns, e 2 Qq and q, can be determined
directly from the spectrum. The energies of the levels in Fig. 14-6 are given by
equations (14-9) and (14-10). The energies of v,, v , and v, are found by the
differences in the energies of the levels. For any two transitions, the resulting
equations can be solved for the two unknowns e 2 Qq and q. It is interesting to
point out that for a nucleus with I = 1 in an axially symmetric field, only one
line is expected in the spectrum [see Fig. 14-6(A)].
E = K(1 + n
FIGURE 14-6 Energy levels
for / = 1 under different
M -+ conditions. (A) q = 0 and
applied magnetic field
E = K(1 -i) Ho = 0. (B) q # 0 and H 0 - 0.
(C) q # 0 and Ho # 0 but
constant.
m =0 E =-2K
(A) (B) (C)
The energies of the quadrupole levels as a function of t have been

calculated 2 ) for cases other than I = 1 or '/2. Tables have also been compiled')
that permit calculations of q and e 2 Qq from spectral data for nuclei with I = /2,
7/2, and '/2. When q is appreciable, the selection rule Am = + 1 breaks down,
and spectra containing Am = 2 bands are often obtained.
14-3 EFFECT OF A MAGNETIC FIELD ON THE SPECTRA
When an nqr experiment is carried out on a sample placed in a static magnetic

field, the Hamiltonian fl describing the influence of the magnetic field on the
nuclear magnetic dipole moment must be added to the nqr Hamiltonian:
Hm = -gN#NH -I (14-11)
In weak magnetic fields (gN/NH < e2 Qq) the magnetic field acts as a perturbation
on HQ. In general, the influence of this term on the energies is to shift the energy
of a non-degenerate quadrupole level and split a doubly degenerate level [i.e., one
having m # 0; see equation (14-6)]. This change in energy of the non-degenerate

quadrupolar levels is indicated in Fig. 14-6(B) and (C) for a nucleus in which
I = 1 and q # 0.
For a nucleus with I = 1, a quadrupole spectrum with two lines can arise
from either of two distinct situations: for P # 0 as noted above (Fig. 14-6(B)), or
for nuclei with il = 0 located in two non-equivalent lattice sites. Examination
of the spectrum of the sample in an external field allows a distinction to be made
between these two possibilities. In the former case (q # 0), two lines would again
be observed, but with different energies than those obtained in the absence of the
field; in the latter case, each doubly degenerate level would be split, giving a
spectrum with four lines.
As mentioned earlier, the levels E+1 /2 and E, 32 are each doubly degenerate
for I = 32 in a non-symmetric field. As a result, e 2 Qq and q cannot be determined
directly. This degeneracy is removed by a magnetic field, with four levels resulting.
Four transitions are observed in the spectrum: +1/2 - +3/2, +1/2 - -3/2,
-/2 - +3/2, and - /2- _3/2 (Alm| = +1). The energy differences corre-
sponding to these transitions are functions of H, e 2 Qq, and q [see equations
(14-7) and (14-8)], so q and q can be evaluated(7) for this system from the spectra
of the sample taken both with and without an applied magnetic field.
14-4 RELATIONSHIP BETWEEN ELECTRIC FIELD

GRADIENT AND MOLECULAR STRUCTURE
Our next concern is how we obtain information about the electronic structure
of a molecule from the values of q and q. The field gradient at atom A in a
molecule, qmotA, and the electronic wave function are related by equation (14-2):
A
qm0 = e Y [ZB(3 cOS 2 OAB - 1)/RAB3 ] _ 0 [(3 cos 2 OAn - 1)/rA3If
(14-12)
The molecular field gradiant qm.,is seen to be a sensitive measure of the electronic
charge density in the immediate vicinity of the nucleus because equation (14-12)
involves the expectation value (1/r 3 >. The first term in the equation is a
summation over all nuclei external to the quadrupolar nucleus, and the second
term is a summation over all electrons. If the molecular structure is known, the
first term is readily evaluated. ZB is the nuclear charge of any atom in the
molecule other than A, the one whose field gradient is being investigated; OAB is
the angle between the bond axis or highest-fold rotation axis for A and the radius
vector from A to B, RAB. The second term represents the contribution to the
field gradient in the molecule from the electron density, and it is referred to as
the electric field gradient q,,. Finally, 0 is the ground state wave function, and
OAn is the angle between the bond or principal axis and the radius vector rA. to
the nth electron. This integral is difficult to evaluate. In the LCAO approxima-
tion, we can write
q1 A iu j Ji* (14-13)
qel e2U i. j f(
14-4 Relationship Between Electric Field Gradient and Molecular Structure 613
where 4A = (3 cos 2 0 _ )r 3 ; u is the index over the molecular orbitals; and i

and j are indices for atomic orbitals. Cju and Cju represent the LCAO coefficients
for the atomic orbitals (pi and -pj in the u molecular orbital. The integral may
involve one, two, or three centers. Obviously, good wave functions are required,
and an involved evaluation of equation (14-13) is needed to interpret q. For
certain atoms (e.g., N and Cl), the three-center contribution is small and can be
ignored. Separating the one- and two-center terms in equation (14-13) and
abbreviating the intergrals f qidAqp dr as qA'j, etc., we can write equation (14-13)
as (14-14):
S A CA 1
q1 jA= -e[2 Z ~.qAii
i 2 + 4Z Y Y
BuA
Cju Cjuq~i + 2 Y
B#A
Y Ciu2 Aj (14-14)
U i u i j u j _
There have been several semi-empirical methods proposed for evaluating

electric field gradients. Cotton and Harris") assumed that the nuclear term of
equation (14-12) (i.e., the first term) was cancelled by the part of qeI^ arising from
the gross atomic populations on the neighbor atom B; i.e.,
B#A
[B3COS2LAB
RAB
3 - 1) e[2 Y
u
B
B#A
Y
j
ju2qAj + 2Y
u
A B#A
B Ciu ju A
(14-15)
This leads to
mA YA Ci ue22
q + 2 Y2 A B#A
CuCjuqA -16)
u i u i j
Further, by assuming that the two-center integral can be formulated as propor-

tional to the overlap integral
qA" = SjgAq" (14-17)

we can write:
A B#A
-
Amos eY qi2 Ciu2 + 2Cju Sij (14-18)
i u U j
We see then that the field gradient is given by multiplying qAj by the gross
atomic population, P [equation (3-31)], or:
A 3cos2 _ I
m -eP ( 3 qi (14-19)
For an atom A with valence s and p orbitals, the above summation is over four
orbitals for which qA for the s orbital is zero, and
q = -eq (Pz - -Px - P

where qat is the field gradient for one electron in a p orbital. The quantity in
parentheses in equation (14-18) is just the molecular-orbital expression for the
gross atomic population P of atomic orbital cpi. (See Chapter 3.) Depending on
the relative populations of p., px, and p,, q can be positive or negative. If the
quadrupole coupling constant is expressed in megahertz, we have
( 2 hqX ) P - P) - P (14-20)
2
where e 2 Qq/h is the quadrupole coupling constant, and e aQqa/his the coupling
constant for a single electron in the p orbital. The p. orbital should be coincident
with q., in order to apply these equations with much accuracy.
Equation (14-20) is the molecular orbital analog to the valence-bond
expression of Townes and Dailey,(9 ) which had been reported earlier. This
approach is based on the following arguments. Since the s orbital is spherically
symmetric, electron density in this orbital will not give rise to a field gradient. As
long as the atom being studied is not the least electronegative in the bonds to
other atoms in the molecule, the maximum field gradient at this atom, qmoi, is
the atomic field gradient, qal, for a single electron in a p. orbital of the isolated
atom. When the atom being investigated is more electronegative than the atom
to which it is bonded, the quadrupolar atom has greater electron density around
it in the molecule than in the isolated atom. The relationship between the electron
2
"occupation" of the p orbitals of a quadrupolar atom in a molecule, e Qqat, and
the quantity e2 Qqmoi (which is determined from the nqr spectrum of the molecule
under consideration) is given by:
e 2Qq0 mo = [1 - s + d - i(1 - s - d)]e 2Qqal (14-21)
where e2 Qqat is the quadrupole coupling constant for occupancy of the p orbital
by a single electron, s is the fraction of s character employed by the atom in the
bond to its neighbor, d is the fraction of d character in this bond, and i is the
fraction of ionic character in the bond (for a molecule A-B, 4' = COA + do, and
2
i = C2 - d ). When the atom being studied is electropositive, i changes sign.
5
Values of e2 Qq3a and qat have been tabulated for several atoms.' 1) When
-10,
7E bonding is possible, this effect must also be included. A modified form of
2
equation (14-21) has also been employed: ( )
e2Qq0 mo = (1 - s + d - i - n)e 2 Qqat (14-22)
where 7Tis the extent of 7c bonding, and all other quantities are the same as before.
As the amount of ionic character in a bond increases, the electronic
2
environment approaches spherical symmetry (where qmo, = 0) and e Qqmoi de-
creases. Hybridization of the p orbital with an s orbital also decreases e 2 Qqmol,
as indicated by equation (14-22). Mixing of the s orbital with the p orbital
decreases the field gradient, because the s orbital is spherically symmetric. In a
covalent molecule, d orbital contribution to the bonding increases the field
gradient.
There are several problems associated with these approaches to the inter-
pretation of field gradients. First, even if the above approach were correct for
the system of interest, there are four unknowns in equation (14-22) and we have
14-4 Relationship Between Electric Field Gradient and Molecular Structure 615
only one measurable quantity, e2 Qq..j. Investigators attempting to interpret this
quantity are thus forced to assume the answer and provide a reasonable
interpretation of the data based on these models. Second, O'Konski and Ha(' 3 )
have shown that the assumption of Cotton and Harris(') indicated in equation
(14-15) is not generally correct. When this equation does not apply, equations
(14-20), (14-21), and (14-22) are also not correct.
A semiempirical approach to the interpretation of field gradients has been
reported," 4 ) which does not make the suspect approximations described above.
The reader is referred to the original literature for details of this method.
Electrostatic, E, covalent, C, and transfer, T, parameters derived from bond
14
energies are reported to provide an interpretation of chlorine field gradients.( b)
One additional factor complicating the quantum mechanical calculation and
interpretation of nqr and Mbssbauer (vide infra) parameters is the Sternheimer
effect.' 5) This effect is the polarization of the originally spherically symmetric
inner shell electrons by the valence electrons. When the core electrons lose their
spherical symmetry, they contribute to the field gradient at the nucleus. This
effect, like spin polarization (discussed in Chapter 12), is an artifact of not
performing a full molecular orbital calculation on the whole crystal. There are
two contributions to Sternheimer shielding, and these can be illustrated if we
consider the ligands or ions external to the electron density of the atom as point
charges. The spherical expansion of the core electron density is illustrated for
positive point charges in Fig. 14-7(A) and (B). The asymmetry induced is shown
in Fig. 14-7(C). We have artificially broken up the polarization to illustrate these
two contributions to the Sternheimer shielding.
FIGURE 14-7 Schematic

illustration of the
Sternheimer effect. (A) A
spherical shell of electron
density; (B) the radial
expansion resulting from two
external positive charges;
(C) the elliptical polarization
+ 0by two positive charges.
(A) (B) (C)
The shielding for closed shell systems is usually expressed by the equation:
e2 Qq0 b. (1
0 - )e2Qqo
where e 2 Qq, is the field gradient calculated in the absence of any Sternheimer
effects and yr is the shielding parameter for the ith charge located at a distance
r from the nucleus. The sign of y,, is a function of the distance. For charges
external to the core electron density of the central atom, y,i will generally be
negative and is said to give rise to an antishielding contribution. For charges
inside the valence orbitals of the central atom, the sign of y,, is generally positive
and gives rise to a shielding contribution. Shielding constants have been calcu-
lated for charges external to a large number of ions and are designated by the
symbol ym. The value of yr used to evaluate the Sternheimer effect in molecules
is difficult to determine but appears to be considerably below that of ym. Most
workers in the field assume that this effect is constant in a similar series of
complexes.
14-5 APPLICATIONS
NQR spectra of a number of molecules containing the following nuclei have been
reported: 2 7 Al, 7 5As, 19 7 Au, ' 0 B, 1"B, 135 Ba, 137 Ba, 2 09Bi, 7 9 Br, 8 1Br, 4 3Ca,
35
Cl, 3 7Cl, 5 9Co, 6 3Cu, 6 9Ga, 7 t Ga, 2 H, 2 0 1Hg, 1271, "5 In, 2 5 Mg, 14 N, 5 Mn,
23Na,
93
Nb, 170, 18 5Re, 187Re, 33S, "iSb, "3Sb, 18'Ta.
The interpretation of e2Qq Data

As mentioned above, it is not possible to interpret e 2Qq values rigorously in
terms of ionic character, 7t bonding, and s and d hybridization. For compounds
in which there are large differences in ionic character, the effect can be clearly
seen, as indicated by the data in Table 14-1. More positive e2 Qqmo, values are
obtained for ionic compounds. Attempts have been made to "explain" e2 Qq data
for a large number of halogen compounds by assuming that d hybridization is
not important and estimating ionic character or s hybridization.(12,16)
When the differences in ionic character are large enough to be predictable,
the e2Qqmoj values manifest the proper trends. However, for systems in which
the differences in ionic character are not obvious from electronegativity and other
considerations, the interpretation of e2 Qq differences is usually ambiguous in
terms of the relative importance of the effects in equation (14-22). One of the
more successful studies involves a series of substituted chlorobenzenes. A linear
relation is found between the Hammett u constant of the substituent and
the quadrupole resonance frequency. 7 ) The data obeyed the equation v
(MHz) = 34.826 + 1.024a. The electron-releasing substituents have negative o-
values and give rise to smaller e 2 Qq values, because the increased ionic character
in the C-Cl bond makes the chlorine more negatively charged.
TABLE 14-1. Values of e 2OQmo, for Some Diatomic

Halides
2 2
Molecule' e 0q.m Molecule" e 0q...
FCl -146.0 BrCI 876.8

BrCl -103.6 LiBr 37.2
ICl -82.5 NaBr 58
TICI -15.8 DBr 533
KCI 0.04 FBr 1089
RbC1 0.774 DI -1827
CsC1 3 Nal -259.87
5
"Values for e 2 Qq., are 1 C = 109.74, "Br = 769.76, and 1271=
2
-2292.84. e Qq.m applies to the underlined atom.
The general problem of covalency in metal-ligand bonds in a series of

complexes has been investigated by direct measurement of the quadrupole
resonance.(1 1 By comparing the trends in force constants with the changes in
chlorine quadrupole coupling constants, a consistent interpretation of the data
for a whole series of MCl"- compounds has been offered."' A rigorous
interpretation of the data (especially where small differences are involved) is again
hampered by the lack of information regarding the four variables in equation
(14-22).
Two transitions are observed in the nqr spectra of 55Mn (I = 5/2), so both
q., and rqcan be obtained. Measurements on a series of organometallic manga-
nese carbonyls have been reported. ) Substitution of methyl groups in the series
(CH 3)nC6H6 -,Mn(CO) 3 causes only minor changes in the asymmetry param-
eter. A summary of cobalt-59 nqr investigations has appeared.(") The magni-
tudes of the field gradients are indicative of cis or trans octahedal geometry. The
e2Qq value of [cis-Co(en)2 Cl2]Cl is 33.71, whereas that for [trans-Co(en)2Cl 2 ]Cl
is 60.63. A rationalization of this difference will be provided in the next chapter,
in the section on partial field gradient parameters.
Thermochemical bond energies have been estimated and compared with the
results from chlorine-35 nqr for the GeCl6 2 , SnCl6 2 and PbCl 2 ~ ions. The
authors(2 2) support an earlier conclusion(2 3 ) that these results are not consistent
with assigning a higher electronegativity to lead than to germanium or tin.
This coverage of quadrupole investigations concerned with elucidating
covalency and other characteristics of chemical bonds is by no means complete.
The presentation is brief because of the ambiguity that exists in the interpretation
of the results. More information than just the field gradient is needed to sort out
the variables in equation (14-22).
Effects of the Crystal Lattice on the Magnitude of e2Oq

A further complication and limitation of the application of nqr spectroscopy
results from the fact that direct measurement of nuclear quadrupole transitions
can be obtained only on solids. For very complex molecules, this is the only
source of nuclear quadrupole information because of the complexity of the
microwave spectrum. Measurements on solids introduce the complexities of
lattice effects. For those molecules which have been studied by both methods
(direct measurement and microwave), it is found that e 2 Qq is usually 10 to 15%
lower in the solid state. It has been proposed(5) that the decrease is due to
increased ionicity in the solid. There are several examples of molecules being
extensively associated in the solid but not in the gas phase or in solution (e.g.,
12 and CNCl). In these cases, considerable care must be exercised in deducing
molecular properties from e 2 Qq values. The e2 Qq values for cobalt in selected
cobalt complexes illustrates 2 1 ) the problem. The values for trans-
[Co(NH 3 )4 Cl 2]Cl and trans-[Co(en) 2 Cl21]C1 are 59.23 MHz and 60.63 MHz; i.e.,
there is a difference of 1.40 MHz. On the other hand, trans-[Co(en)2 C12 ]NO 3
has an e 2 Qq(Co) value of 62.78 MHz. Changing the anion causes a difference of
2.15 MHz. The e 2 Qq(Cl) values of K 2 SnCl6 and [(CH 3 )4 N] 2 SnCl6 are
15.063 MHz and 16.674 MHz, respectively. Potential causes for these differences
have been offered.( 2 2b, 2 2 c) Examples in which the crystal lattice affects the
number of lines observed in a spectrum will be discussed in the next section.
Structural Information from NQR Spectra

Since different field gradients will exist for non-equivalent nuclei in a molecule,
we should expect to obtain a different line (or set of lines, depending on I) for
each type of nuclear environment. In general, the environment of an atom as
determined by nqr studies is in agreement with results obtained from x-ray
studies. Only one line is found"') in the halogen nqr spectrum of each of the
following: K2 SeCl 6 , Cs2 SeBr 6, (NH4 )2 TeCl,, (NH 4 )2 SnBr,, and K2 PtCl,. This
is consistent with 0 structures for these anions.
As mentioned above, the following effects can give rise to multiple lines in
the nqr spectrum:
1. Chemically non-equivalent atoms in the molecule.
2. Chemically equivalent atoms in a molecule occupying non-equivalent
positions in the crystal lattice of the solid.
3. Splitting of the degeneracy of quadrupole energy levels by the asym-
metry of the field gradient. Splittings of the quadrupole levels by other magnetic
nuclei in the molecule, similar to spin-spin splittings in nmr, are often not detected
in nqr spectra, but are being found with improved instrumentation. Usually these
splittings are less than, or of the same order of magnitude as, the line widths.
One case in which such splitting has been reported(24 ) is in the spectrum of H10 3 -
The 127I nucleus is split by the proton.
The non-equivalence of lattice positions (2) is illustrated in the bromine nqr
spectrum of K 2SeBr,, which gives a single line at room temperature and two
lines at dry ice temperature. A crystalline phase change accounts for the
difference. NQR is one of the most powerful techniques for detecting phase
changes and obtaining structural information about the phase transitions.
Although K 2PtI 6 , K 2 SnBr 6 , and K 2TeBr, contain "octahedal" anions, the
halides are not equivalent in the solid lattice, and the halide nqr spectra all consist
of three lines, indicating at least three different halide environments.
Four resonance lines were found in the chlorine quadrupole spectra of each
of TiC 4,(2 a SiCl 4 ,(2 sb) and SnCl 4 .(2 sb> It was concluded that the crystal struc-
tures of these materials are similar.
The problem of distinguishing non-equivalent positions in the lattice from
chemical non-equivalence in the molecular configuration (which is related to the
structure of the molecule in the gas phase or in a non-coordinating solvent) is
difficult in some cases. In general, the frequency difference for lines resulting from
non-equivalent lattice positions is small compared to the differences encountered
for chemically non-equivalent nuclei in a molecule. When only slight separations
between the spectral lines are observed, it is difficult to determine from this
technique alone which of the two effects is operative. Another difficulty en-
countered in interpretation is illustrated by the chlorine spectrum of M2 Cl 0
species.(2 6) Even though the Nb 2 C11 molecule has a structure containing both
terminal and bridging chlorines, only a single chlorine resonance assigned(21) to
the bridging atom is observed in the spectrum. In Re 2 Clio and W 2 Clio, a large
number of resonances were obtained.(2 1 ) Consequently, arguments based on
assigning the number of different types of nuclei in the molecular structure to
the number of resonances observed can be of doubtful validity.
NQR studies of 4 N (I = 1) are difficult to carry out but produce very
interesting results. Since I = 1, both e 2 Qq and the asymmetry parameter rqcan be
evaluated from the nqr spectrum. In BrCN, the "N nqr resonance is a doublet.
This could result from two non-equivalent nitrogen atoms in the crystal lattice
or from a splitting of the nitrogen resonance because of an asymmetric field
gradient. The former explanation was eliminated by a single crystal x-ray
study.!" The structure of solid BrCN was found to consist of linear chains of
the type:
Br-C---N| - Br--CEN --.-.-Br-C--NI
The nitrogen has axial symmetry here, and only one line is expected. However,
it is proposed that interactions between chains reduce the symmetry at the
nitrogen and lead to the two lines. Various resonance forms can be written for
BrCN, and the e2 Qq values indicate that the bromine has a formal positive
charge. An appreciable increase in e2 Qq for bromine is observed in the solid
relative to the gaseous state spectrum of BrCN. This could be due to increased
contributions to the ground state from the structure Br*CN- in the solid because
of stabilization of Br' by coordination. If the N .... Br-C bond is described as
a pd hybrid, the increased d contribution in the bromine carbon bond will also
increase e2 Qq.
The "N quadrupole transitions in pyridine and various coordination
compounds of pyridine show(2") large changes in the transition energies. As
discussed in conjunction with Fig. 14-6, three transitions (vt, vo, and v_) are
expected when i # 0, and e2 Qq and q can be determined from the data. Typical
results are summarized in Table 14-2.
The iodine nqr spectrum of solid iodine indicates a large asymmetry param-
eter, q.(29) Since the iodine atom in an iodine molecule is axially symmetric,
the large asymmetry is taken as indication of intermolecular bonding in the solid.
A large il in the iodine nqr spectrum of the molecule H10 3 supports the structure
10 2 (OH) instead of H10 3 . The structure H10 3 has a C3 axis, so qx= q,,.
References 1 and 2 contain many additional examples of studies of this kind.
Information regarding 7r bonding can be obtained from the asymmetry
parameter q. Methods for evaluating q for various nuclei have been discussed.
A single a bond to a halogen should give rise to an axially symmetric field
gradient. Double bonding leads to asymmetry, and the extent of 7[ bonding is
TABLE 14-2. 14 N Quadrupole Transitions and Field Gradient

Parameters in Pyridine and Coordinated Pyridine (All Frequen-
cies in kHz, Temp. = 77 K). (Crystallographic Non-Equivalences,
Where Present, Are Resolved.)
e2Qq
Compound v, v_ vo h
Pyridine (Py) 3892 2984 908 4584 0.396

Pyridinium nitrate 1000 580 420 1053 0.798
Py 2 ZnCl 2 2387 2078 309 2977 0.207
2332 2038 294 2913 0.202
Py 2Zn(NO3) 2 2124 1884 240 2672 0.180
2097 1877 220 2649 0.166
Py 2 CdCl2 2850 2298 552 3432 0.320
related to rj. It was concluded that there is appreciable (- 5%) carbon-halogen

nr bonding in vinyl chloride, vinyl bromide, and vinyl iodide.!") Bersohn(") has
made a complete study of the problem of quantitatively estimating the extent of
carbon-halogen 71 bonding from rq.
The values of rqobtained from the nqr spectra of SiI 4 , Gel 4 , and SnI 4 were
interpreted to indicate a very small degree (about 1%) of double bond char-
acter13 1) in the halogen bond to the central atom. This could be due to a
solid state effect.
The 7 7 As, 1 2 1 , 123 Sb, 2 09 Bi, 35 3 7 Cl, and 7 9 -81Br nqr spectra of compounds
with general formula R 3 MX 2 have been studied 33 ) for R = CH 3 , CH 2C6 H,, and
CH,, for X = F, Cl, and Br, and for M = As, Sb, and Bi. Very small asymmetry
parameters were found, indicating a threefold axis in all compounds. The results
suggest that most of the compounds are trigonal bipyramidal. However, the
arsenic compound [(CH 3)3AsBr*]Br- is not, but probably has a cation with
C3, symmetry.
The nqr spectra (3 7 Cl, 35 Cl, 2 1 Sb, 2 7I) of 21Cl - AlCl 3 and 21C - SbCl,
indicate(3 4) that these materials should be formulated as ICl 2 'AlCl 4 - and
4
ICl2 +SbCl,-. The nqr spectra of the v-shaped cations ICl2 ', 13', and 12 C1
have been studied in several different compounds. The 3 5Cl nqr spectra of the
chloroaluminate group in a wide variety of M,(AlCl 4 ), compounds have been
studied.(3"> The transition energies can be used to indicate whether the relatively
free ion AICl4- exists or whether this anion is strongly coordinated to the cation.
"Ionic" AlCl 4 transitions are found in the 10.6 to 11.3 MHz range. Strong
coordination of AlCl 4 - results in an elongation of the Al-Cl bridge bonds and
an appreciable increase in the range and average frequency of the chlorine
transitions.
In the next chapter, on M6ssbauer spectroscopy, we shall discuss an additive,
partialfieldgradient (pfg), model for correlating field gradients at central atoms.
It is useful for deducing structures of molecules from experimental M6ssbauer
data. This model can also be used to infer structures from nqr data. Since most
of the data used to derive and test the pfg model are M6ssbauer results, we shall
treat this topic in the next chapter.
14-6 DOUBLE RESONANCE TECHNIQUES
Nuclear double resonance techniques have been reported(3-3s ) for observing

quadrupolar transitions. These methods greatly increase sensitivity over previous
techniques and also permit one to observe transitions with very low frequencies;
39
e.g., in 2 H spectra, transitions are observed in the 100 to 160 kHz region. )
The adiabatic demagnetization double resonance experiment is a novel
technique in this general category. Consider a sample that has a quadrupolar
nucleus in a molecule containing several protons. When this sample is placed in
a magnetic field and we wait long enough for equilibrium to be obtained, there
will be an excess of proton nuclear moments aligned with the field that undergo
Larmor precession and give rise to a net magnetization, as discussed in the
chapters on nmr. When the sample is removed from the field, the net magnetiza-
tion is reduced to zero as the sample is removed because the individual moments
now become aligned with their local fields. A random orientation of these local
fields in the absence of an external field produces a zero net magnetization. This
14-4 Double Resonance Techniques 621
is illustrated in Fig. 14-8 on the left-hand side, in the region labeled "sample
removed."
If T, is longer than the time required to remove the sample, the excess
population of + '/2 spins will remain, but they now precess about a net local
field felt at the nucleus from spin-spin interaction with neighboring protons. Over
the full sample, the magnetization is zero; but when this sample is reinserted into
a magnetic field, magnetization is simultaneously induced into the sample
without having to wait the required time for the T 1 process. This is illustrated
in the region of Fig. 14-8 labeled "sample reinserted." The intensity of the
magnetization can be measured by employing a 90 pulse immediately after
reinserting the sample in the magnetic field and measuring the FID curve (see
Fig. 14-8). If the time between the removal of the sample from the magnetic field
and reinserting it is long compared to T, the magnetization will decrease as the
spins become randomized.
JL
FIGURE 14-8 Plot of the
M magnetization when a
sample is removed from a
magnetic field, reinserted,
subjected to a 90 pulse, and
then allowed to undergo free
induction decay.
0
0 Time | | FID
t << T
Sample Sample 900
Removed Reinserted Pulse
Now consider an experiment in which the sample is irradiated with an rf

frequency corresponding to the quadrupolar nucleus B transition after the sample
has been removed from the field. Furthermore, we shall assume that the time
between removal and reinsertion is small compared to T, for the protons. The
effect of this rf field is to randomize the B nucleus by inducing quadrupolar
transitions in the B spin system. Provided that the appropriate conditions are
met, in terms of the amplitude of the applied rf field in relationship to the local
field experienced at the protons, the randomization of the B spin system
influences that of the proton system. This occurs by the following process. When
the sample is removed from the field, the energy difference between the m = + '/2
and -'/2 states (i.e., the transition energy of the H nucleus) decreases toward
zero. In the process, there is a time at which the energy difference for the hydrogen
nuclei matches the energy difference of the quadrupolar states of the B nucleus.
A resonance energy exchange occurs, tending to randomize the proton nuclei.
The process of randomization is often referred to as an increase in the spin
temperature of the system. As a result of the randomization of the proton system
from transitions in B, the magnetization that is recovered when the sample is
returned to the magnetic field is less than it would otherwise be. As a result, the
FID measured is less than it would otherwise be. When the frequency used for
the quadrupole transition is not appropriate for resonance, the B nucleus is not
randomized and a large amount of the magnetization is recovered. By system-
atically incrementing the frequency of the B transmitter, the "spectra" of the
quadrupole transitions of the B system are mapped out in terms of their effect
on the FID of the abundant proton spin system.
By using a spin echo double resonance experiment,(3 9 -41) much of the
inhomogeneous dipolar broadening (crystal imperfections, etc.) that leads to very
broad lines in the direct nqr experiment can be eliminated. The nqr of Al 2 Br,
has been determined with this technique.(40 )
If one has two quadrupolar nuclei surrounded by nuclei with I = 0 [e.g.,
as in D-Mn(Co)j, the dipolar coupling of the manganese and deuterium nuclei
can be observed.(42 ) As discussed in the nmr chapters, the bond distance can be
obtained from the magnitude of the dipolar coupling. A Mn-D bond distance
of 1.61 A is calculated from the 55Mn nqr spectrum, in excellent agreement with
the neutron diffraction result.(4)
REFERENCES 1. C. P. Slichter, "Principles of Magnetic Resonance," Harper & Row, New York, 1963,
CITED Sec. 6.3.
2. R. Bersohn, J. Chem. Phys., 20, 1505 (1952).
3. R. Livingston and H. Zeldes, "Tables of Eigenvalues for Pure Quadrupole Spectra,"
Oak Ridge Natl. Lab. Rept. ORNL-1913 (1955).
4. a. T. P. Das and E. L. Hahn, "Nuclear Quadrupole Coupling Spectroscopy,"
b. E. A. C. Lucken, "Nuclear Quarupole Coupling Constants," Academic Press, New
York, 1969.
5. C. T. O'Konski, "Determination of Organic Structures by Physical Methods," Vol.
2, ed. F. C. Nachod and W. D. Phillips, Academic Press, New York, 1962.
6. J. A.Pople, W. G. Schneider, and H. J. Bernstein, "High Resolution Nuclear Magnetic
Resonance," McGraw-Hill, New York 1959.
7. C. Dean, Phys. Rev., 86, 607A (1952).
8. F. A. Cotton and C. B. Harris, Proc. Nati. Acad. Sci. U.S., 56,12 (1966); Inorg. Chem.,
6, 376 (1967).
9. C. H. Townes and B. P. Dailey, J. Chem. Phys., 17, 782 (1949).
10. M. H. Cohen and F. Reif, "Solid State Physics," Vol. 5, ed. F. Seitz and D. Turnbull,
11. W. J. Orville-Thomas, Quart. Rev., 11, 162 (1957).
12. B. P. Dailey, J. Chem. Phys., 33, 1641 (1960).
13. C. T. O'Konski and T. K. Ha, J. Chem. Phys., 49, 5354 (1968).
14. a. W. D. White and R. S. Drago, J. Chem. Phys., 52, 4717 (1970).
b. R. S. Drago, N. Wong, and D. Ferris, J. Amer. Chem. Soc., 113, 1970 (1991).
15. R. M. Sternheimer, Phys. Rev., 130, 1423 (1963); ibid., 146, 140 (1966) and references
therein.
16. C. H. Townes and B. P. Dailey, J. Chem. Phys., 23, 118 (1955); W. Gordy, Disc.
Faraday Soc., 19, 14 (1955); M. A. Whitehead and H. H. Jaffe, Trans. Faraday Soc.,
57, 1854 (1961).
17. H. C. Meal, J. Amer. Chem. Soc., 74, 6121 (1952); P. J. Bray and R. G. Barnes, J.
Chem. Phys., 27, 551 (1957) and references therein.
18. D. Nakamura, K. Ito, and M. Kubo, Inorg. Chem., 1, 592 (1962); ibid., 2, 61 (1963);
J. Amer. Chem. Soc., 82 5783 (1960); ibid., 83, 4526 (1961); ibid., 84, 163 (1962); M.
Kubo and D. Nakamura, Adv. Inorg. Chem. Radiochem., 8, 257 (1966).
19. T. L. Brown, W. G. McDugle, and L. G. Kent, J. Amer. Chem. Soc., 92, 3645 (1970).
20. T. B. Brill and A. J. Kotlar, Inorg. Chem., 13, 470 (1974).
21. T. L. Brown, Accts. Chem. Res., 408, (1974) and references therein.
22. a. W. A. Welsh, T. B. Brill, et al., Inorg. Chem., 13, 1797 (1974).
b. T. B. Brill, R. C. Gearhart, and W. A. Welsh, J. Mag. Res., 13, 27 (1974).
c. T. B. Brill, J. Chem. Phys., 61, 424 (1974).
23. R. S. Drago and N. A. Matwiyoff, J. Organometal. Chem., 3, 62 (1965).
24. R. Livingston and H. Zeldes, J. Chem. Phys., 26, 351 (1957).
25. a. A. H. Reddoch, J. Chem. Phys., 35, 1085 (1961).
b. A. L. Schawlow, J. Chem. Phys., 22, 1211 (1954).
26. P. A. Edwards and R. E. McCarley, Inorg. Chem., 12, 900 (1973).
27. S. Geller and A. L. Schawlow, J. Chem. Phys., 23, 779 (1955).
28. Y. N. Hsieh, G. R. Rubenacker, C. P. Cheng, and T. L. Brown, J. Amer. Chem. Soc.,
99, 1384 (1977).
29. H. G. Dehmelt, Naturwissenschaften, 37, 398 (1950).
30. J. A. Howe and J. H. Goldstein, J. Chem. Phys., 26, 7 (1957); 27, 831 (1957) and
references therein.
31. R. Bersohn, J. Chem. Phys., 22, 2078 (1954).
32. H. Robinson, H. G. Dehmelt, and W. Gordy, J. Chem. Phys., 22, 511 (1954); S. Kojima
et al., J. Phys. Soc. Japan, 9, 805 (1954).
33. T. B. Brill and G. G. Long, Inorg, Chem., 13, 1980 (1970).
34. D. J. Merryman and J. D. Corbett, Inorg. Chem., 13, 1258 (1974)
35. D. J. Merryman, P. A. Edwards, J. D. Corbett, and R. E. McCarley, Inorg. Chem.,
13, 1471 (1974).
36. S R. Hartmann and E. L. Hahn, Phys. Rev., 128, 2042 (1962).
37. R. E. Slusher and E. L. Hahn, Phys. Rev., 166, 332 (1968).
38. D. T. Edmonds et al., "Advances in Quadrupole Resonance," Vol. 1, p. 145, Heyden,
London, 1974; Rev. Pure Appl. Chem., 40, 193 (1974).
39. J. L. Ragle and K. L. Sherk, J. Chem. Phys., 50, 3553 (1969).
40. M. Emshwiller, E. L. Hahn, and D. Kaplan, Phys. Rev., 118, 414 (1960).
41. J. L. Ragle et al., J. Chem. Phys., 61, 429, 3184 (1974).
42. a. P. S. Ireland, L. W. Olson, and T. L. Brown, J. Amer. Chem. Soc., 97, 3548 (1975).
b. P. S. Ireland and T. L. Brown, J. Mag. Res., 20, 300 (1975).
43. S. J. LaPlaca et al., Inorg. Chem., 8, 1928 (1969).
M. H. Cohen and F. Reif, "Solid State Physics," Vol. 5, ed. F. Seitz and D. Turnbull, ADDITIONAL
Academic Press, New York, 1957. REFERENCES
T. P. Das and E. L. Hahn, "Nuclear Quadrupole Resonance Spectroscopy," Academic
E. A. C. Lucken, "Nuclear Quadrupole Coupling Constants," Academic Press, New York,
1969.
J. A. S. Smith, ed., "Advances in Nuclear Quadrupole Resonance," Heyden and Sons,
London: Vol. 1, 1974; Vol. 2, 1975; Vol. 3, 1977.
EXERCISES 1.a. Calculate the energies of all the quadrupolar energy states for a nucleus with I = 2.
Express the energies as a function of e2 Qq.
b. How many transitions are expected, and what is the relationship between the
energy of the transitions and e2 Qq?
2. a. Using the equations presented in this chapter for the energies of the 0 and
+ 1 levels of a nucleus with I = 1 in an asymmetric field, calculate the frequency
2
in terms of e Qq and rqfor the 0 -* + 1 and 0 -+ 1 transitions.
b. Express the energy difference between the two transitions in part a in terms of rq
and e 2Qq.
c. Show how rqand q can be determined from this information.
3. Describe an nqr experiment that would give information regarding the extent of 7E
bonding in the phosphorus-sulfur bonds in PSCl 3 and (CH 5 )3 PS. Can you determine
whether the sulfur is hybridized sp 2 and utilizes a p orbital in bonding or whether
the p. and p, orbitals of sulfur participate equally in bonding with nqr experiments?
(Note: for 33S, I = 3/2-)
4. It has been reported that the 1271 quadrupole resonance in AsI 3 is a singlet but has
a very large asymmetry parameter.(28 ) A single crystal x-ray study indicates that the
As is nearly octahedral. Explain the large asymmetry parameter.
5. Indicate the number of resonance lines expected for the following nuclei under the
conditions given:
a. 1271 (I = 5/ 2 ); r 0; H, = 0.
b. 14N (I = 1); rq 0; Ho = 0.
c. 7 'As (I = 3/2); r;= 0; Ho = 0.
21
d. Sb (I = 5/2 ); = 1; Ho = 0.
e. "N (I = 1); r/=1; Ho =0.
f "N (I= 1); =1; H# 0.
6. The quadrupolar energy of a nucleus is given by
e2 Qq[3m2 - I(I + 1)]

41(21- 1)
s"Co has I = 7/2 and a natural abundance of 100%.
a. How many cobalt nqr transitions will be observed for K 3 Co(CN)6 ? What are the
transition energies in terms of e2 Qq?
b. Repeat part a for K 3Co(CN),Br.

Exercises 625
7. Consider the nitrogen nqr of each of the following systems. How many lines would
you expect with and without a magnetic field? (For 14 N, I = 1.)
a. NH 3
b. NH 4 *
c.:NQ
8. The 59Co (1 = /2) frequencies in ClSnCo(CO)4 occur at 35.02 MHz (+2 - i2),
23.37 MHz (3/2 - ±/2), and 11.68 MHz (±'/2 - ±3/2). Calculate q and e2 Qq.
(Hint: What are the ratios of the frequencies when q = 0?)
9. A compound having the formula CH 3InI 2 is known. It is believed to be an ionic

compound, (CH 3)2 In*Inl 4 -. Provided there are no crystallographically non-equiv-
alent cations or anions, how many resonance lines would you expect for .. 5in and
1 27
I in this compound? The cation has q = 0.05. Which structure of this cation is
suggested by this small value?
10. In solid pyridine (CHN) at 77 K, 14N lines are found at 3.90 and 2.95 MHz.
What are e 2Qq/h and q for nitrogen in pyridine?
35
11. The CI lines in the spectrum of HgCl 2 lie at 22.05 and 22.25 MHz at 300 K.
22.05 22.25
3
In HgCl 2 - dioxane, a single 1C1 line at 20.50 MHz is found at 300 K. In the dioxanate
complex, an Hg +- O donor-acceptor interaction occurs.

a. What is the probable source of the line splitting in pure HgCl 2 ?
b. The electric quadrupole moments of 3 sCl and 37 Cl are 0.079 x 10 -24 cm2 and
0.062 x 10 -24 cm 2 , respectively. At what frequency would you expect to find 37CI
resonances in HgCl 2 ?
c. In terms of the p orbital populations of equation (14-13), rationalize the decrease

in the 3 sCl resonance frequency in the dioxanate adduct of HgCl2 compared to
that in pure HgCl2.
15Spectroscopy
Mossbauer
15-1 INTRODUCTION
M6ssbauer spectroscopy,O' abbreviated as MB spectroscopy, involves nuclear

transitions that result from the absorption of 7-rays by the sample. This transition
is characterized by a change in the nuclear spin quantum number, I. The
conditions for absorption depend upon the electron density about the nucleus,
and the number of peaks obtained is related to the symmetry of the compound.
As a result, structural information can be obtained. Many of the concepts and
symbols used in this chapter have been previously discussed in Chapter 14.
To understand the principles of this method, first consider a gaseous system
consisting of a radioactive source of y-rays and the sample, which can absorb
7,-rays. When a gamma ray is emitted by the source nucleus, it decays to the
ground state. The energies of the emitted y-rays, E,, have a range of 10 to 100
keV and are given by equation (15-1):
E, = E, + D - R (15-1)
where E, is the difference in energy between the excited state and ground state
of the source nucleus; D, the Doppler shift, is due to the translational motion of
the nucleus, and R is the recoil energy of the nucleus. The recoil energy, similar
to that occurring when a bullet leaves a gun, is generally 10-2 to 10-3 eV and
is given by the equation:
E 2
(15-2)
R = /2mc
where m is the mass of the nucleus and c is the velocity of light. The Doppler
shift accounts for the fact that the energy of a 7-ray emitted from a nucleus in a
gas molecule moving in the same direction as the emitted ray is different from
the energy of a -- ray from a nucleus in a gas molecule moving in the opposite
direction. The distribution of energies resulting from the translational motion of
the source nuclei in many directions is referred to as Doppler broadening. The
left-hand curve of Fig. 15-1 represents the distribution of energies of emitted
y-rays, E,, from this effect. The breadth of the curve results from Doppler
broadening. The dashed line in Fig. 15-1 is taken as E,, the energy difference
626 between the nuclear ground and excited states of the source. The energy difference,
Source Sample
|- R R
E FIGURE 15-1 Distribution of

energies of emitted and
absorbed y-rays.
cc
Energy
R, between the dotted line and the average energy of the left-hand curve is the
recoil energy transmitted to the source nucleus when a /-ray is emitted.
In MB spectroscopy the energy of the y-ray absorbed for a transition in the
sample is given by:
E, = E, + D + R (15-3)
In this case, R is added because the exciting 7-ray must have energy necessary
to bring about the transition and effect recoil of the absorbing nucleus. The
quantity D has the same significance as before, and the value of E, is assumed
to be the same for the source and the sample. The curve in the right half of Fig.
15-1 shows the distribution of 7-ray energies necessary for absorption. The
relationship of the sample and source energies can be seen from the entire figure.
As indicated by the shaded region, there is only a very slight probability that the
,-ray energy from the source will match that required for absorption by the
sample. Since the nuclear energy levels are quantized, there is accordingly a very
low probability that the /-ray from the source will be absorbed to give a nuclear
transition in the sample. The main cause for nonmatching of 7,-ray energies is
the recoil energy, with the distribution for emission centered about E, - R,
whereas that for absorption is centered about E, + R. The quantity R for a
gaseous molecule (~ 10 eV) is very much larger than the typical Doppler energy.
The source would have to move with a velocity of 2 x 104 cm sec- t o obtain
a Doppler effect large enough to make the source and sample peaks overlap, and
these velocities are not readily obtainable. However, if the quantity R could be
reduced, or if conditions for a recoilless transition could be found, the sample
would have a higher probability of absorbing 7-rays from the source. As indicated
by equation (15-2), R can be decreased by increasing m, the mass. By placing the
nucleus of the sample and source in a solid, the mass is effectively that of the
solid and the recoil energy will be small as indicated by equation (15-2). For this
reason, MB spectra are almost always obtained on solid samples employing solid
sources.
628 Chapter 15 M6ssbauer Spectroscopy
By placing the source and sample in solid lattices, we have not effected
recoilless transitions for all nuclei, but we have increased the probability of a
recoilless transition. The reason for this is that the energy of the y-ray may cause
excitation of lattice vibrational modes. This energy term would function in the
same way as the recoil energy in the gas; i.e., it would decrease the energy of the
emitted particle and increase the energy required for absorption. Certain crystal
properties and experimental conditions for emission or absorption will leave the
lattice in its initial vibrational state; i.e., conditions for a recoilless transition will
be satisfied. It should be emphasized that these conditions simply determine the
intensity of the peaks obtained, for it is only the number of particles with matching
energy that is determined by this effect. We shall not be concerned with the
absolute intensity of a band, so this aspect of MB spectroscopy will not be
discussed. It should be mentioned, however, that for some materials (usually
molecular solids), lattice and molecular vibrational modes are excited to such an
extent that very few recoilless transitions occur at room temperature and no
spectrum is obtained. Frequently, the spectrum can be obtained by lowering the
temperature of the sample.
By going to the solid state we have very much reduced the widths of the
resonance lines over that shown in Fig. 15-1. The Doppler broadening is now
negligible, and R becomes -10 - eV for a 100 keV gamma ray and an emitting
mass number of 100. The full width of a resonance line at half height is given by
the Heisenberg uncertainty principle as AE = h/T = 4.56 x 10-16/0.977 x 10 -7 =
4.67 x 10-9 eV or 0.097 mm sec-1 (for 57 Fe). The line widths are infinitesimal
compared to the source energy of 1.4 x 104 eV. The range of excited state lifetimes
for Mbssbauer nuclei is ~10 - sec to 10- " sec, and this leads to line widths of
10-" eV to 10-6 eV for most nuclei. This subject is treated in references 1 to
5, which contain a more detailed discussion of the entire subject of MB
spectroscopy.
Our main concern will be with the factors affecting the energy required for
y-ray absorption by the sample. There are three main types of interaction of the
nuclei with the chemical environment that result in small changes in the energy
required for absorption: (1) resonance line shifts from changes in electron
environment, (2) quadrupole interactions, and (3) magnetic interactions. These
effects give us information of chemical significance and will be our prime concern.
Before discussing these factors, it is best to describe the procedure for
obtaining spectra and to illustrate a typical MB spectrum. The electron
environment about the nucleus influences the energy of the y-ray necessary to
cause the nucleartransition from the ground to excited state, i.e., Er in the sample.
The energy of y-rays from the source can be varied over the range of the energy
differences arising from electron environments in different samples by moving
the source relative to the sample. The higher the velocity at which the source is
moved toward the sample, the higher the average energy of the emitted y-ray (by
the Doppler effect) and vice versa. The energy change AE, of a photon associated
with the source moving relative to the sample is given by:
AE, = v EY cos 0 (15-4)

c
where E, is the stationary energy of the photon, v, is the velocity of the source,
and 0 is the angle between the velocity of the source and the line connecting the
source and the sample. When the source is moving directly toward the sample,
cos 0 = 1. In order to obtain an MB spectrum, the source is moved relative to
the sample, and the source velocity at which maximum absorption of y-rays
occurs is determined.
3
Consider, as a simple example, the MB spectrum of Fe +Fe"(CN)6 [where
Fe ± and Fe"' designate weak and strong field iron(III), respectively]. This
3
substance contains iron in two different chemical environments, and y-rays of

two different energies are required to cause transitions in the different nuclei. To
obtain the MB spectrum, the source is moved relative to the fixed sample, and
the absorption of y-rays is plotted as a function of source velocity as shown in
Fig. 15-2. The peaks correspond to source velocities at which maximum y-ray
FIGURE 15-2 MB spectrum of

FeFe(CN) 6 .
E
z
I I I
-1.0 0 1.0
mm sec-1
absorption by the sample occurs. Negative relative velocities correspond to
moving the source away from the sample, and positive relative velocities
correspond to moving the source toward the sample. The relative velocity at
which the source is being moved is plotted along the abscissa of Fig. 15-2, and
this quantity is related to the energy of the y-rays. For a "Fe source emitting a
14.4 keV y-ray, the energy is changed by 4.8 x 10 -8 eV or 0.0011 Cal mole -1 for
every mm sec -' of velocity imposed upon the source. This result can be calculated
from equation (15-4):
AE 1 mm sec-' x 14.4 x 103 eV = 4.80 x 10-' eV

3.00 x 10" mm sec
This energy is equivalent to a frequency of 11.6 MHz (v = E/h, where

h = 4.14 x 10 eV sec). For other nuclei having a y-ray energy of E7 (in keV),
E
1 mm sec =11.6 x ' MHz
14.4
Referring again to the abscissa of Fig. 15-2, one sees that the energy difference
between the nuclear transitions for Fe 3 + and Fe"' in FeFe(CN), is very small,
corresponding to about 2 x 10-8 eV. The peak in the spectrum in Fig. 15-2 at
0.03 mm sec-1 is assigned(6 ) to Fe"' and that at 0.53 to, the cation Fe 3 + by
comparison of this spectrum with those for a large number of cyanide complexes
of iron. Different line positions that result from different chemical environments
are indicated by the values for the source velocity in units of cm -1 or mm sec -1,
and are referred to as isomer shifts, center shifts, or chemical shifts. We shall now
proceed with a discussion of the information contained in the parameters obtained
from the spectrum.
15-2 INTERPRETATION OF ISOMER SHIFTS
The two different peaks in Fig. 15-2 arise from the isomer shift differences of the
two different iron atoms in octahedral sites. The isomer shift results from the
electrostatic interaction of the charge distribution in the nucleus with the electron
density that has a finite probability of existing at the nucleus. Only s electrons
have a finite probability of overlapping the nuclear density, so the isomer shift
can be evaluated by considering this interaction. It should be remembered that
p, d, and other electron densities can influence s electron density by screening
the s density from the nuclear charge. Assuming the nucleus to be a uniformly
charged sphere of radius R and the s electron density over the nucleus to be a
constant given by 0,2(0), the difference between the electrostatic interaction of a
spherical distribution of electron density with a point nucleus and that for a
nucleus with radius R is given by
E = K[0,2f(0)] R 2 (15-5)
where K is a nuclear constant. Since R will have different values for the ground
state and the excited state, the electron density at the nucleus will interact
differently with the two states and thus will influence the energy of the transition;
i.e.,
6E, - 6Eg = K[0, 2 (O)](Re2 - Rg2 ) (15-6a)
where the subscript e refers to the excited state and g to the ground state. The
influence of , 2 (0) on the energy of the transition is illustrated in Fig. 15-3 for
5
Fe, which has I = '/2 for the ground state and I = /2 for the excited state. The
energies of these two states are affected differently by 0,2(0), and the transition
energy is changed.
FIGURE 15-3 Changes in the (e)

energy of the M6ssbauer E 3/ (e)
transition for different values of N
iS2(0). This is a graphical E
illustration of equation (15-16a) R
with two different values of G (g)
i2(O) for an 57Fe2 nucleus. The G
differences in 0 S (0) must result
from a cubic or spherical /
/= (g)
distribution of bonded atoms in
order for this diagram to apply. ' 2
(0) C 4, 2(0)
The R values are constant for a given nucleus, but 0, 2 (o) varies from
compound to compound. The center shift in the M6ssbauer spectrum is the
difference between the energy of this transition in the sample (or absorber) and
the energy of the same transition in the source. This difference is given by the
difference in equations of the form of (15-6a) for the source and sample, or:
C.S. = K(Re 2 R 2 ){[0r2 (o)]a _ [01s2(0)]} (1

( 5-6b)
15-3 Quadrupole Interactions 631
where the subscripts a and b refer to absorber and source, respectively. Standard
sources are usually employed (e.g., "7 Co in Pd for iron Mossbauer spectra, or
BaSnO 3 for tin spectra). The ' 7 Co decays to 57 Fe in an excited state via electron
capture. The excited 5 7 Fe decays to stable 57 Fe by y-ray emission. When standard
sources are employed [f,2 (O)], is replaced by a constant, C. Furthermore, the
change in radius R, - R9 is very small, leading to the following commonly
employed expression for the center shift:
C.S. =K' [R1[0s2(q _ C] (15-6c)

R
where 6R = R - Rg; C is a constant characteristic of the source; and K' is 2KR 2

Both K' and 6R/R are constants for a given nucleus, so the center shift is directly
proportional to the s electron density at the sample nucleus. The term center shift
is used for the experimentally determined center of the peak; the term isomer
shift is now used when the center shift has been corrected for the small Doppler
contribution from the thermal motion of the M6ssbauer atom. The sign of 6R
depends upon the difference between the effective nuclear charge radius, R, of
the excited and ground states (R, 2 - Rg 2 ). For the 57 Fe nucleus, the excited state
is smaller than the ground state, and an increase in s-electron density produces
a negative shift. In tin, the sign of 6R is positive, so the opposite trend of shift
with s electron density is observed.
As mentioned above, electron density in p or d orbitals can screen the electron
density from the nuclear charge by virtue of the fact that the electron density in
d and p orbitals penetrates the s orbital. Hartree-Fock calculations show (6,7) that
a decrease in the number of d electrons causes a marked increase in the total s
electron density at the iron nucleus. Accordingly, with comparable ligands and
with negative 3R/R, Fe2 + has an appreciably larger center shift than Fe3". When
these ions are examined in a series of molecules, the interpretation becomes more
difficult, for the d, s and p electron densities are modified by covalent bonding.
For 57 Fe, for example, an increase in 4s density decreases the center shift, while
an increase in 3d density increases the center shift. A series of high spin iron
complexes have been interpreted on this basis.(71 In the case of .. 'Sn, the center
shift increases with an increased s electron density and decreases with an increase
in p electron density.(8)
15-3 QUADRUPOLE INTERACTIONS
The discussion of the center shift in the previous section applies to systems with
a spherical or cubic distribution of electron density. As discussed in Chapter 14,
the degeneracy of nuclear energy levels for nuclei with I > i/2 is removed by a
non-cubic electron or ligand distribution. For non-integral spins, the splitting
does not remove the + or - degeneracy of the m, levels, but we obtain a different
level for each + m, set. Thus, the electric field gradient can lead to I + '/2 different
levels for half-integer values of I (e.g., two for I = 3/2 corresponding to + 1/2 and
± 2). For integer values of I we obtain 21 + 1 levels (e.g., five for I = 2
corresponding to 2, 1, 0, -1, -2). The influence of this splitting on the nuclear
energy levels and the spectral appearance is illustrated(9') in Fig. 15-4 for 57 Fe.
The ground state is not split but the excited state is split, leading to two peaks
m
3
I2
Q.S.
2 2 Z
0
0
(n
CO Q.S.
Center Quadrupole
Shift Splitting i 1 _I _ _ _1__ _ _
2 -1 ' 0 1 -4.0 0 4.0
Center Shift mm sec-1
mm sec-1
(A) (B) (C)
FIGURE 15-4 The influence of a non-cubic electronic environment on (A) the
nuclear energy states of 57Fe and (B) the M6ssbauer spectrum. (C) The iron MB
spectrum of Fe(CO) 5 at liquid N2 temperature.(9)
in the spectrum. The center shift is determined from the center of the two resulting
peaks. When both the ground and excited states have large values for I, complex
M6ssbauer spectra result.
The Hamiltonian for the quadrupole coupling is the same as that discussed
for nqr.
e 2Qq -(+ + (2 + Z)
H 41(2I - 1) [(3 1) + (q/2) ± 2
For the I = 3/2 case ( 7 Fe and "'Sn), the quadrupole splitting Q.S. is given by
Q.S. = - e2 Qq(1 + 2 /3)1 2 (15-7)

2
The symbols have all been defined in the nqr chapter. For 57 Fe, q and q cannot
be determined from the quadrupole splitting. The sign of the quadrupole coupling
constant is another quantity of interest. If m, = /2 is at high energy, the sign
is positive; the sign is negative if /2 from I = 3/2 is highest. From powder
spectra, the intensities of the transitions to + 1/2 and + 3/2 are similar, and it
becomes difficult to determine the sign. The sign can be obtained from spectra
of ordered systems or from measurement of a polycrystalline sample in a magnetic
field (vide infra). For systems in which the I values of the ground and excited
states are larger than those for iron, the spectra are more complex and contain
more information. The splitting of the excited state will not occur in a spherically
symmetric or cubic field but will occur only when there is a field gradient at the
nucleus caused by asymmetric p or d electron distribution in the compound. A
field gradient exists in the trigonal bipyramidal molecule iron pentacarbonyl, so
a splitting of the nuclear excited state is expected, giving rise to a doublet in the
spectrum as indicated in Fig. 15-4(C).
If the t2g set and the eg set of orbitals in octahedral transition metal ion
complexes have equal populations in the component orbitals, the quadrupole
splitting will be zero. Low spin iron(II) complexes (t2,') will not give rise to
15-4 Paramagnetic Mbssbauer Spectra 633
quadrupole splitting unless the degeneracy is removed, and these orbitals can
interact differently with the ligand molecular orbitals. On the other hand, high
spin iron(II) (t2 g4 eg) has an imbalance in the t2g set, and a large quadrupole
splitting is often seen. If the ligand environment about iron(II) were perfectly
octahedral, d,y, d,, and d. would be degenerate and no splitting would be
detected. However, this system is subject to Jahn-Teller distortion, which can
lead to a large field gradient. When the energy separaton of the t2, orbitals from
Jahn-Teller effects is of the order of magnitude of kT, a very temperature-
dependent quadrupole splitting is observed. The ground state in the distorted
complex can be obtained if the sign of q is known. The sign can be obtained
from oriented systems or from studies in a large magnetic field. Similar
considerations apply to high spin and low spin iron(III) compounds.
The factors contributing to the magnitude of the field gradient were discussed
in Chapter 14. It was shown there that these data were of limited utility in
providing further information about bond types.
15-4 PARAMAGNETIC MOSSBAUER SPECTRA
As shown in Fig. 15-4, a non-cubic electronic environment in an iron complex

splits the degenerate I = 3/2 excited state of 7 Fe into m1 states 1/2 and i/2-
The M6ssbauer spectrum consists of a doublet corresponding to the two
transitions shown in Fig. 15-4. When an effective magnetic field, H,,,, acts on
this system, the degeneracies of the ± '/2 ground state as well as the + /2
excited states are removed"" as shown in Fig. 15-5. The ordering of the levels
reflects the fact that the ground-state moment is positive and the excited state
moment is negative. The resulting six-line spectrum is shown in Fig. 15-5(B).
Since the transitions are magnetic dipole in character, the selection rule is Am = 0,
+ 1 as shown in Fig. 15-5(A). The Hamiltonian is
H, = gNNI- H, (15-8)
where gN = 0.18 and I for the nuclear ground state and g = -0.10 and
I = '/2 for the excited state. The intensity of the lines changes with 0, the
orientation of the Heff relative to the direction of the M6ssbauer radiation. For
an ordered sample, the Am = 0 transition has zero intensity for a field parallel
to the radiation. When 0 = 90 , the Am = 0 transitions are the most intense of
the six.
The effective field of equation (15-8) consists of the internal field and the
applied field:
Hef= H + Hint (15-9)
and is given by
Hi = S) - A/gNN (15-10)
where KS> is the expectation value of the spin discussed in Chapters 11 and 12
and A the hyperfine coupling tensor. The internal magnetic field controls the
magnetic hyperfine and it depends in a complex way on the zero field splitting
634 Chapter 15 Mdssbauer Spectroscopy
m,
63
+2
35
+2
3 E
2 a4
FIGURE 15-5
Magnetic and quadrupole splitting 1
in a ferromagnetic 57Fe compound.
(A) Energy level diagram. (B) 9E,
Expected M6ssbauer spectrum.
[Copyright @ 1973 McGraw-Hill Magnetic Quadrupole
Book Co. (UK) Limited. From G. M. splitting splitting
Bancroft, "M6ssbauer (A)
Spectroscopy." Reproduced by
permission.]
4
Y 2 5
1 6
-Ve 0 +Ve
Velocity (mm s-)
(B)
parameters D and E, on the electronic g tensors and A tensors, and on the

orientation of the molecule relative to the applied field.
Internal fields exist in some systems in the absence of an applied magnetic
field. The KS> of equation (15-10) is zero for diamagnetic systems and for systems
with integer values of S in the absence of an applied field."' These spectra consist
of quadrupole doublets. For componds with half-integer spin (Kramers' doublets),
two cases need be considered. With fast electron relaxation, the KS> averaged
over all thermally accessible states is zero in zero magnetic field, leading to
quadrupole doublets. In the slow relaxation limit, magnetic M6ssbauer spectra
can be obtained, generally at liquid helium temperatures. These spectra generally
are studied at weak applied field to simplify them by decoupling hyperfine
interactions with ligand nuclei.
In applied fields, magnetic M6ssbauer splittings can be observed for both
integer and half-integer spin systems. The applied field mixes the states producing
a non-zero expectation value of the spin. The resulting spectra are interpreted
using the spin Hamiltonian formalism" 0 , given in equation (15-11).
f = D [,2 _ S(S + 1) + E(5/ - 52) + #5-gf + - A-
2
gN flN + e2 Qq 1 12 _(±1) + ( _2) EFG
4I(2I - 1)
(15-1 1)
By standing the spectra as a function of temperature, field strength, and field

direction, the various quantities in the spin Hamiltonian can be determined. For
Kramers' doublets complimentary information can be obtained from the epr.
However for integer S systems, e.g., high spin iron(II), magnetic M6ssbauer
provides the sole means of obtaining electronic structural details."" Reference
14 presents applications of magnetic M6ssbauer.
15-5 MOSSBAUER EMISSION SPECTROSCOPY

7
s"Co decays by an electron capture to 5 7 Fe (T, for " Fe is 0.1 psec), populating
an excited state of the iron nucleus. The emitted y-rays can be absorbed by a
7
standard single-line absorber to investigate the energy levels of the " Fe nuclei
produced when the source decays. The cobalt-57 analogue of the compound to
be studied is prepared and used as the source. Information regarding the
short-lived iron complex in the source is obtained"') from this experiment. One
must be sure that the desired iron complex remains intact when the high energy
iron atoms are formed in the cobalt decay process. The results obtained from
some oxygenated complexes of 57 Co protoporphyrin IX dimethyl ester as a
function of the axial base attached are shown in Table 15-1. These are to be
compared with values of AEQ and J for oxygenated hemoglobin, obtained by
absorption measurements, of 2.23 qnd 0.27 respectively.
2
TABLE 15-1. M6ssbauer Emission Studies. ( )
(Quadrupole splittings and isometric shifts for oxygenated
complexes of 5 7Co protoporphyrin IX dimethyl ester. The
ligands coordinated trans to dioxygen are listed in the first
column.)
Ligand AEG(mm/sec) 6Fe(Mm/SeC)
1-methyl imidazole 2.17 0.29

1,2-dimethyl imidazole 2.32 0.30
pyridine 2.28 0.27
piperidine 2.25 0.30
ethylmethyl sulfide 2.27 0.30
Fig. 15-6(A) shows the emission spectrum of the five-coordinate 1-methyl

imidazole complex before oxygenation; Fig. 15-6(B) shows the spectrum of the
same complex after oxygenation.
This technique is particularly important when the parent iron compound is
difficult to prepare and isolate.
15-6 APPLICATIONS
A few chemical applications of M6ssbauer spectroscopy have been selected for

discussion that are illustrative of the kind of information that can be obtained.
Table 15-2 summarizes pertinent information about isotopes that have been
studied by this technique.
Facsimiles of spectra obtained on some iron complexes are given in Fig.
15-7. As mentioned previously, for high spin iron complexes in which all six
2 - (A)
3 I
FIGURE 15-6 Emission

M6ssbauer spectra of (A) the
1-methyl imidazole adduct of0
cobalt protoporphyrin IX
dimethyl ester and (B) its 02
adduct.
2 - (B
3-
I I | | I
3 2 1 0 -1 -2 -3
Absorbervelocity (mm/so)
ligands are equivalent, a virtually spherical electric field at the nucleus is expected
for Fe3 +(d5) (t2 3 e,') but not for Fe2 +(d6 )(t2 ge'g2 ). As a result of the field gradients
at the nucleus, quadrupole splitting should be detected in the spectra of high spin
iron(II) complexes but not for high spin iron(III) complexes. This is borne out
in spectra A and B of the complexes illustrated in Fig. 15-7. For low spin
complexes, iron(II) has a configuration t2 6 and iron(III) has t,.. As a result,
quadrupole splitting is now expected for iron(III) but not iron(II) in the strong
field complexes. This conclusion is confirmed experimentally by the spectra of
ferrocyanide and ferricyanide ions. When the ligand arrangement in a strong field
iron(II) complex does not consist of six equivalent ligands, e.g., [Fe(CN),NH, 3]3-
quadrupole splitting of the strong field iron(II) will result. The quadrupole splitting
is roughly related to the differences in the d orbital populations (see Chapter 3) by
qve = Kd -Ndz + Nda, , + Nda, - 2 (Nd + Nd)]

TABLE 15-2. M6ssbauer Isotopes of Chemical Interesta
Gamma Natural SR
energy Half life' Half width O od abundance 104-
Isotope (keV) of precursor (mm sec-1) Spin (Barns)
1
(10~ ' cm ) 2
(M) R
57 1 3
Fe 14.4 270 d 0.19 - - +0.3 23.5 2.19 -18 +4
2 2
1 3 5
6 Ni 67.4 1.7 h; 3.3 h 0.78 2 2 0.13 7.21 1.19
2 2
6 7
Zn 93.3 60 h; 78 h 3.13 x 10- -> - +0.17 1.18 4.11
2 2
73
9 7
Ge 67.0 76 d 2.19 - -0.26 3.54 7.76
2 2
9 7
83Kr 9.3 83 d, 2.4 h 0.20 - - +0.44 18.9 11.55 +4±2
2 2
99 5 3
Ru 90 16.1 d 0.15 - - >0.15 1.42 12.72
2 2
07 1 7
1 Ag 93.1 6.6 h 6.68 x 10 -" - 0.54 51.35
2 2
1 3
"t 9 Sn 23.9 245 d 0.62 - - -0.07 13.8 8.58 +3.3 + 1
2 2
5 7
"2 Sb 37.2 76 y 2.10 - - -0.29 2.04 57.25 -8.5 + 3
2 2
5 1 3
12 Te 35.5 58 d 4.94 - - +0.19 2.72 6.99 + 1
2 2
7 5
1291 27.8 33 d; 70 m 0.63 - - -0.55 3.97 0 +3
2 2
29 1 3
Xe 39.6 1.6 x 107 y 6.84 - - -0.41 1.95 26.44 0.3
2 2
7 5
"3CS 81.0 7.2 y 0.54 - - -0.003 1.02 100
2 2
7 9
1??Hf 113.0 6.7 d, 56 h 4.66 - - +3 1.20 18.50
2 2
7 9
181Ta 6.25 140 d, 45 d 6.48 x 10 - - +4.2 17.2 99.99
2 2
182w 100.1 115 d 2.00 0 ->2 -1.87 2.46 26.41 +1.3
5 7
187Re 134.2 23.8 h 203.8 - - +2.6 0.54 62.93
2 2
1860s 137.2 90 h 2.37 0 - 2 +1.54 2.89 1.64
93
3 1
1 Ir 73.1 32 h 0.59 2 -> - +1.5 0.30 62.7 +0.6
2 2
1 3
195Pt 98.7 183 d 17.30 - - - 0.63 33.8
2 2
3 1
197Au 77.3 65 h, 20 h 1.85 - -> +0.58 0.44 100 +3
2 2
"Copyright C 1973 McGraw-Hill, (UK) Limited. From G. M. Bancroft, "Mossbauer Spectroscopy." Reproduced by permission.
'd = days, h = hours, y = years, m = minutes.
'The ground-state quadrupole moment, where both ground and excited states have 1 > 2
'Cross-section for absorption of a M6ssbauer gamma ray.
638 Chapter 15 Mssbauer Spectroscopy
FIGURE 15-7 M6ssbauer

spectra of some iron(II) -0.2 0 +0.2 +04
and iron(III) complexes.
-0.1 0 +0.2
(A) Spin-free
(A) (B)
iron(Il)-FeSO4 - 7H 20.
(B) Spin-free iron(Ill)-FeCl3 .
(C) Spin-paired
iron(II)-K 4Fe(CN), - 31H2 0.
(D) Spin-paired
iron(Ill)-K 3Fe(CN) 6 . [From P.
R. Brady, P. P. F. Wigley,
and J. F. Duncan, Rev. Pure
Appl. Chem., 12, 181 (1962).] | I I I
-0.2 -0.1 0 +QI +0.2

(D)
Here, qvaience is the contribution to q from valence electrons in the d orbitals. For
p-electrons we have
qence, = K,[ - N, + 2 (NP, + NP)]
TABLE 15-3. Quadrupole Splitting, AEo, and Isomer Shift, 6, for Some Iron
Compounds (6 and AEo in mm sec')
Compound AEo 6 Compound AEo S
High Spin Fe(ll) Low Spin Fe(II)
FeSO 4 -7H 20 3.2 1.19 K 4 [Fe(CN) 6 ] 3H 2 0 0.13
3.15 1.3 0.16
FeSO 4 (anhydrous) 2.7 1.2 <0.1 +0.05
Fe(NH 4 )2 (SO 4 )2 -6H 2 0 1.75 1.19 Na 4 [Fe(CN) 6 ] 10H2 0 <0.2 1.01
1.75 1.3 Na 3 [Fe(CN) 5 NH,] 0.6 0.05
FeC12 -4H 2 0 3.00 1.35 K 2 [Fe(CN) 5 N0] 1.85 0.27
FeC 4 H 4 0 6 2.6 1.25 1.76 0.28
FeF 2 2.68 - Zn[Fe(CN) 5 N] 1.90 0.27
FeC 2 O4 -2H 2 O 1.7 1.25
High Spin Fe(iII) Low Spin Fe(ill)

K3 [Fe(CN)]3 - -0.12
FeC13 6H< 0.2 0.85 0 +0.17
FeCN3 (anhydrous) 0.2 0.5 0.26 -0.15
FeCI3 2NH 4 C- H 2 0 0.3 0.45 Na 3 [Fe(CN)NH 0.60 -0.17
Fe(N0 3 ) 3 20 0.4 0.4 K e)9H
Fe 2 (C2 0 4 ) 3 0.5 0.45
Fe 2 (CH 4 0 6 )3 0.77 0.43
Fe 2 3 0.12 0.47
Values measured at room temperature for AE, and the isomer shift, 3, for
a number of iron complexes have been collectedm) and are listed in Table 15-3.
For iron complexes, isomer shifts in a positive direction correspond to a decrease
in electron density in the region of the nucleus. For high spin complexes, a
correlation exists between isomer shift and s electron density. An increase in 3
of 0.2 mm sec-' is equivalent to a decrease in charge density of 8% at the
nucleus.(!) The negative values obtained for the low spin ferricyanides compared
to high spin iron(III) complexes indicate more electron density at the nucleus in
the ferricyanide ions. This has been explained as being due to extensive 7r bonding
in the ferricyanides, which removes d electron density from the metal ion, which
in turn decreases the shielding of the s electrons. This effect increases s electron
density at the nucleus and decreases 3. Both strong a donors and strong 7n
acceptors decrease 3.
The MB spectrum of the material prepared from iron(III) sulfate and
K 4Fe(CN), is identical to the spectra for the compounds prepared either from
iron(II) sulfate and K 3 Fe(CN), or by atmospheric oxidation of the compound
from iron(II) sulfate and K 4 Fe(CN),. The spectra of these materials indicate that
the cation is high spin iron(III), while the anion is low spin iron(II).
The MB spectrum of sodium nitroprusside, Na 2 Fe(CN),NO, has been
investigated. 9 ) This material has been formulated earlier as iron(II) and NO'
because the complex is diamagnetic. The MB spectrum consists of a doublet with
a AEQ value of 1.76 mm sec - and a 3 value of -0.165 mm sec '. Comparison
of this value with reported results(8) on a series of iron complexes led the authors
to conclude that the iron 3 value is close to that of iron(IV). The magnetism and
MB spectrum are consistent with a structure in which there is extensive 7t bonding
between the odd electron in the t 2 set of the orbitals of iron and the odd electron
on nitrogen, as illustrated in Fig. 15-8. The filled 7cbonding orbital would need
a large contribution from the nitrogen atomic orbital, and the empty 7Eantibonding
orbital would have a larger contribution from the iron atomic orbital to produce
iron(IV). More of the 71 electron density would be localized on nitrogen and the
3 value for iron would approach that of iron(IV) because of decreased shielding
of the s electrons by the d electrons. Since electron density is being placed in
what was previously a 7r antibonding orbital of nitric oxide, a decrease in the
N-O infrared stretching frequency is observed. The very large quadrupole FIGURE 15-8 Orbitals
splitting is consistent with very extensive 7r bonding in the Fe-N-O link. involved in the Fe-N 7t
M6ssbauer spectra are often useful in determining the oxidation state of bonding to the NO group in
atoms. It has been shown that the spectra originally assigned to some high spin [Fe(CN)5 (NO)] 2-
iron(II) chelates of salicylaldoxime and other chelates were, in effect, those of an
iron(III) oxidation product. The oxidized materials gave a center shift [relative
to Na 2Fe(CN),NO - 2H 20] of ~-0.6 mm sec -' and a small quadrupole splitting.
This smaller-than-expected quadrupole splitting for iron(II) was rationalized in
terms of it-backbonding. The authentic material prepared under conditions that
were rigorously air-free gave normal iron(II) center shifts and large quadrupole
splittings, ~ 1.4. Values for the center shifts of various oxidation states of high
spin iron compounds are shown in Table 15-4.
The effect of pressure on the M6ssbauer spectra of a wide variety of iron
compounds is reported.( 1 ) Upon application of pressure (- 165 kbar) to trisace-
tylacetonato iron(III), a new species forms that is attributed to iron(II). The
change is reversed upon removal of the pressure.
The determination of the oxidation states of tin compounds from MB spectra
TABLE 15-4. Isomer Shifts for High Spin Iron Compounds [Shifts Relative to
Na2Fe(CN),NO -2H20]
Oxidation state +1 +2 +3 +4 +6
I.S. - +2.2 +1.4 ~+0.7 ~+0.2 -- 0.6
is not as clear-cut as in the case of iron. Values of 6 below 2.65 mm sec -1 are
often due to tin(IV), and those above that value to tin(II). Exceptions are known.
The isomer shifts of some four- and six-coordinate tin(IV) componds vary
considerably as the attached anion is varied.(' 7 )
The MB spectra of the iron pentacarbonyls Fe(CO),, Fe 2 (CO)9 , and
Fe,(CO),, have been reported.( 8"-9) The results are as expected from the known
structures for both Fe(CO), and Fe 2(CO)9 . The structure of Fe,(CO),, deduced
from its MB spectrum was at odds with infrared results and a preliminary x-ray
study. The MB indicated more than one type of iron, as shown in Fig. 15-9(A).
The outer two lines are assigned to one type of iron and the inner two to a
second type. In general, the areas are roughly proportional to the number of a
particular type of iron present. A definitive crystal structure study has supported
the structure shown in Fig. 15-9(B), which was predicted from the M6ssbauer
results. The spectra of Fe(II)X 2 (CO) 2 P 2 (where X = Cl, Br, and I, and
P = phosphines and phosphites) have been interpreted in terms of the five different
isomers that exist.
0
0
I I | I |
-0.5 0 1.0 2.0
Velocity (mm/sec) (B)
(A)
FIGURE 15-9 The M6ssbauer spectrum (A) and structure (B) of Fe3 (CO) .
12
Several systems involving spin equilibrium between high spin and low spin
iron(II) complexes have been studied by M6ssbauer spectroscopy. A typical
result(20 > involves the hexadentate ligand {4-[(6-R)-2-pyridyl]-3-azabutenyl} 3
amine. The spectra obtained when two or three of the R groups are methyl are
characteristic of low spin iron(II) ('A 1) at 77 K, while at 294 K the large isomer
shift and quadrupole splitting found are characteristic of high spin iron(II) (5T).
At intermediate temperatures, both forms are observed in the spectrum. This
establishes that it is a true equilibrium ('T 2 4 'A 1 ) and that these states are
long-lived on the M6ssbauer time scale; i.e., the lifetime must be 10-' sec or
greater.
There are several interesting biological applications of M6ssbauer spectros-
copy.(5i 2 ,21 2, 2) Horseradish peroxidase is an iron(III) heme protein. It can be
oxidized in two one-electron steps, producing red and green compounds. The
M6ssbauer spectrum changes(2 3 ) upon oxidation to either form and in both cases
is interpreted in terms of iron(IV). The removal of the second electron leading
to the green form is believed to come from the protein or porphyrin.
M6ssbauer studies have been of considerable utility in the study of the redox
centers that exist in several classes of iron-sulfur proteins. A crystal structure
investigation of the ferredoxin HP,,d photosynthetic (high potential) protein from
Chromatium has been carried out, and it has been shown to contain an
Fe4 S 4 (SCys) 4 (SCys refers to cystein) core with the cubane-like structure shown Fe'
in Fig. 15-10. Similar units are present in other proteins. The compound HPed
is readily oxidized to form HP... The crystal structure does not provide -RS Fe S
information about the oxidation states of the iron or the charge on the cluster. Fe - -
RS
The M6ssbauer spectra of several proteins containing this unit consist of a
quadrupole split doublet.12 4 1 By comparing the isomer shift in HP,,d with those S Fe,
of other iron systems, it was concluded1 2 41 that the core consists of an average
of two Fe3 and two Fe2 + ions. The irons are antiferromagnetically coupled, FIGURE 15-10 Cubane-like
leading to a diamagnetic material, and the metal electrons in the system are structure of the Fe4S4(S-R-)4
delocalized so that only one kind of iron atom exists. The sensitivity of isomer core of some ferredoxins.
shift to oxidation state is indicated in Table 15-5.
TABLE 15-5. Center Shifts for Various

Iron-Sulfur Compounds
Species" 6
Fe3 + (rubredoxin) 0.25

3Fe3 + lFe2 (Chromatium HP..) 0.32
2Fe 3 + + 2Fe2 + (Chromatium HP,,d 0.42
and ox. ferredoxin)
1Fe 3 + + 3Fe 2 + (red. ferredoxin) 0.57
Fe2+ (rubredoxin) 0.65
"IAn entry such as 3Fe 3 1 + Fe 2 + is meant to imply an
average oxidation state for the iron atoms corresponding
to the given combination.
The model compound [(C2 H,) 4 N] 2 [Fe4 S4 (SCH 2CH,)4 ] has been pre-
pared. 251 The charge on this anion is known to be minus two , so it must be a
2Fe2 + + 2Fe3 + case. The Mossbauer spectrum( 26) shown in Fig. 15-11 and the
crystal structure are similar to those of HP,,d or oxidized ferredoxin with
equivalent iron atoms and an isomer shift of 0.36 mm sec-'. This supports the
assignment( 4 ) of the corresponding ferredoxins as 2Fe 2 + + 2Fe 3 + systems.
The M6ssbauer spectrum of (Et 4 N) 2 [Fe4 S4 (SCH 2C.H 5 )4 ] shown in Fig.
15-11(A) is a quadrupole split doublet arising from the low symmetry about the
iron center. The lines are further split by a magnetic field, as shown in (B). The
solid line in (A) is a least squares fit of the data, employing Lorentzian line shapes.
The solid line in (B) is a computer-generated spectrum employing H = 80
kilogauss, rq= 0, and AEQ = 1.26 mm sec -. The sign of the principal component
of the field gradient is found to be positive from the spectral fit.
FIGURE 15-11 Mssbauer

spectrum of
(Et4N)2[Fe 4S4(SCH 2CH 4)4] (A)
at 1.5 K and Ho = 0, and (B)
at 4.7 K and HO = 80
kOersteds. [Reprinted with
permission from R. H. Holm,
et al., J. Amer. Chem. Soc.,
96, 2644 (1974). Copyright by
the American Chemical
Society.]
-
(B
+2.0 .0 0 -1.0 -2.0

VELOCITY (mm/sec)
As we have mentioned several times, it is difficult to interpret field gradients.

However, it has been found possible to parametrize ions and groups attached to
a central metal ion and to use these parameters, called additive partialquadrupole
splittings, to predict the quadrupole coupling. The basic model is one of a point
charge. In a diagonal electric field gradient coordinate system, the contributions
to V.., V,,, and V, from a charge Z are given by
V- = Zer-3(3 sin 2 0 cos 2 p 1)

V, = Zer -3(3 sin 0 sin p _ 1)
2 2
V2 = Zer- 3(3 cos 2 0 1)
where 0 and p have their usual definitions in the polar coordinate system. For
a tetragonal complex (Fig. 15-12), for example, we must take each ligand (a point
charge) and sum the individual contributions to V., V,, and V.. Table 15-6
contains a summary of this calculation for the cis and trans complexes in Fig.
15-12.
Summing the contributions in Table 15-6, we obtain for the trans complex
V, = V,, = (-2A] + 2[BI)e

V. = (4[A] - 4[B])e
where [A] and [B] are unspecified contributions from ligands A and B,
respectively. For the cis complex we obtain
V. = V, = (A] - [B])e
V = (--2[A] + 2[B])e
z z
B3 , B3
FIGURE 15-12 Geometries
and coordinates for (A) trans
M q) and (B) cis MB4A2-
4B'X
B4
A 'X
A2 B2
(A) (B)
TABLE 15-6. Individual Point Charge Contributions to the EFG Tensor in

trans- and cis-MA 2 B4 .a The Quantity [A] Equals ZAe/rA.
trans-MA 2B4
Ligand
(Fig. 15-12) 0 (p sin 0 cos 0 sin (p cos <p V,,/e V,,Ie V,,e
Al 0 0 0 1 0 1 -[A] -[A] +2[A]

A2 180 0 0 -1 0 1 -[A] -[A] + 2[A]
B 90 0 1 0 0 1 +2[B] -[B] -[B]
B2 90 90 1 0 1 0 -[B] + 2[B] -[B]
B 90 180 1 0 0 -1 +2[B] -[B] -[B]
B4 90 270 1 0 -1 0 -[B] +2[B] -[B]
cis-MA 2B4
Ligand 0 (p sin 0 cos 0 sin p cos p V.,e V,,Ie V,,Ie
A 90 0 1 0 0 1 +2[A] -[A] -[A]

A2 90 90 1 0 1 0 -[A] +2[A] -[A]
B 0 0 0 1 0 1 -[B] -[B] +2[B]
B2 180 0 0 -1 0 1 -[B] -[B] +2[B]
B3 90 180 1 0 0 -1 +2[B] -[B] -[B]
B4 90 270 1 0 -1 0 -[B] +2[B] -[B]
" Copyright C 1973 McGraw-Hill (UK) Limited. From G. M. Bancroft, "M6ssbauer Spectroscopy."
Reproduced by permission.
The ratio of the trans to cis quadrupole splittings is
(4[A] - 4[B])e -2
(-2[A] + 2[B])e
If both isomers are available, M6ssbauer spectra provide a ready means of

distinguishing them, with this ratio. In the partial quadrupole coupling approach,
magnitudes are empirically assigned to [A] and [B], and the field gradients are
predicted. A cis or transgeometry is readily determined with the above equations.

For other geometries, the appropriate equations must be derived. The compound
Fe(CO) 2 (PMe3 )2 12 can exist as five different isomers whose calculated quadrupole
splittings differ considerably. Two isomers can be prepared with observed
quadrupole splittings of -0.90 and + 1.31 mm sec -. These are readily identified
as the (trans P, cis I, cis CO) and all trans isomers, respectively.! 7 ) Partial
quadrupole splitting parameters("' for various ligands bonded to iron(II) are
listed in Table 15-7.
When the values for the ligands in Table 15-7 are used with appropriate
equations for the complexes in Table 15-8, the values listed under the column
labeled "predicted" result.(2 7)
TABLE 15-7. Ligand Partial Quadrupole Splitting (PQS)

2
Parameters( 7) for Iron(II)a
Ligand POS value Ligand POS value
NO * +0.01 P(OPh)3 -0.55

X- -0.30 CO -0.55
N2 -0.37 PPh 2 Et -0.58
N3 -0.38 PPh 2 Me -0.58
CH 3 CN -0.43 depb/2 -0.59
SnC13 -0.43 P(OEt) 3 -0.63
H2 Ob -0.45 depe/2 -0.65
SnPh 3 -0.50 P(OMe) 3 -0.65
NCS- -0.51 PMe 3 -0.66
AsPh 3 -0.51 dmpe/2 -0.70
NH 3b -0.52 ArNC -0.70
NCO_ -0.52 CN-b -0.84
PPh 3 -0.53 H- -1.04
"Copyright C 1973 McGraw-Hill (UK) Limited. From G. M. Bancroft,

"M6ssbauer Spectroscopy." Reproduced by permission.
bPQS values derived from room temperature data.
TABLE 15-8. Predicted and Observed( 2 7 ,28) QS (mm sec -1) at 295 Ka
Observed Predicted
trans-FeCI2 (ArNC)4 +1.55

cis-FeCl 2 (ArNC)4 -0.78 -0.78
[FeCI(ArNC)3]ClO 4 0.73 +0.78
trans-Fe(SnC 3 )2(ArNC) 4 +1.05
cis-Fe(SnCl 3)2(ArNC) 4 0.50 -0.52
cis-FeCISnCl 3(ArNC) 4 0.61 -0.69 (q = 0.60)
trans-FeH 2(depb)2 -1.84
trans-FeH Cl(depe)2 <0.12 -0.20
trans-Fe(EtNC)4 (CN)2 -0.60
cis-Fe(EtNC) 4(CN) 2 0.29 +0.30
trans-[FeH(ArNC)(depe) 2]BPh4 -1.14 -0.98
trans-[FeH(CO)(depe)2]lBPh4 1.00 -0.46
"Copyright c 1973 McGraw-Hill (UK) Limited. From G. M. Bancroft, "M6ssbauer Spectroscopy."

Reproduced by permission.
Other parameters have been reported for use with tin compounds.12 7 ) The
approach has been applied to a large number of systems with a high degree of
success.(2 7 ,2s) There clearly are some shortcomings in the point charge model 2 )
but more work is required to enable one to predict when it will break down in
empirical type applications.
The sign and magnitude of the field gradient can be used to provide
information about the electronic ground state of a transition metal ion complex.
The approximate value of the field gradient for different ground states can be
estimated by adding the d orbital contributions of the populated orbitals,
employing Table 15-9.
TABLE 15-9. Magnitude of q and r for

Various Atomic Orbitals
Orbital q T1
4
P -- (r-'> 0
5
2
pN + - Kr -3> -3
5
2
p, + r-3> +3
5
K>+
(r-3) 0
4
7,- t - 3 0
74
d,+- Kr-3> 0
2
dx -Kr-3> +3
2
dK - r-3) -3
The quantity Kr-'> is the expectation value of 1/r3 for the appropriate
orbital function. This table is constructed by using the various orbital functions
to evaluate the matrix elements:
q-q ' 3 c os2 1
K~3
r
~jq sin2 O cos 29p
In a typical application, the aig2(d.2), e2 a4 (dx,, d,2 ) ground state for ferrocene
is predicted to have a q., value of 2(-4/,(r ->) + 4(4/,(r -3>) = 81<r ->. A large
positive field gradient is observed. In forming the ferricenium cation, a decrease
in the quadrupole splitting is observed, which is consistent with the loss of an
e2, electron. (30)
646 Chapter 15 Mossbauer Spectroscopy
REFERENCES 1. P. G. Debrunner and H. Frauenfelder, in "Introduction to M6ssbauer Spectroscopy,"

CITED ed. L. May, Plenum Press, New York, 1971.
2. G. M. Bancroft, "Mbssbauer Spectroscopy-An Introduction for Inorganic Chemists
and Geochemists,' McGraw-Hill (UK), 1973.
3. M. Weissbluth, Structure and Bonding, 2, 1 (1967).
4. A. J. Bearden and W. R. Dunham, Structure and Bonding, 8, 1 (1970).
5. T. C. Gibb and N. N. Greenwood, "M6ssbauer Spectroscopy," Chapman & Hall, 1971.
6. J. F. Duncan and P. W. R. Wigley, J. Chem. Soc., 1963, 1120.
7. R. E. Watson and A. J. Freeman, Phys. Rev., 120, 1125 (1960).
8. L. R. Walker, G. K. Wertheim, and V. Jaccarino, Phys. Rev. Lett., 6, 98 (1961).
9. L. M. Epstein, J. Chem. Phys., 36, 2731 (1962).
10. a. E. Minck and B. H. Huynh in "ESR and NMR of Paramagnetic Species in
Biological and Related Systems," eds. I. Bertini and R. S. Drago, D. Reidel,
Dordrecht, 1980.
b. E. Munck, Methods Enzymol., 54, 346 (1978).
11. A. Abragam and B. Bleaney, "Electron Paramagnetic Resonance of Transition Ions,"
Chapter 15, Oxford University Press, New York, 1970.
12. a. R. L. Collins and J. C. Travis, M6ssbauer Effect Methodol., 3, 123 (1967).
b. G. Lang, Quart. Rev. Biophys., 3, 1 (1970).
13. a. E. Munck and P. M. Champion, J. Physique, Suppl. 12/35, C6-33 (1974).
b. "M6ssbauer Spectroscopy," Topics in Applied Physics, Volume 5, ed. U. Gonser,
Springer Verlag, New York, 1975.
14. a. R. L. Collins, J. Chem. Phys., 42, 1072 (1965).
b. E. Minck, in "The Porphyrins," ed. D. H. Dolphin, Academic Press, New York,
1975.
c. W. T. Oosterhuis, Structure and Bonding, 20, 59 (1974).
d. G. Lang, Quart. Rev. Biophys., 3, 1 (1970).
15. L. Marchant, M. Sharrock, B. M. Hoffman, and E. Milnck, Proc. Natl. Acad., Sci.
USA, 69, 2396 (1972).
16. A. R. Champion, R. W. Vaughan, and H. G. Drickamer, J. Chem. Phys., 47, 2583
(1967) and references therein.
17. R. V. Parish, Prog. Inorg. Chem., 15, 101 (1972).
18. R. H. Herber, W. R. Kingston, and G. K. Wertheim, Inorg. Chem., 2, 153 (1963).
19. R. H. Herber, R. B. Kingston, and G. K. Wertheim, Inorg. Chem., 3, 101 (1964).
20. M. A. Hoselton, L. J. Wilson, and R. S. Drago, J. Amer. Chem. Soc., 97, 1722 (1975).
21. W. Marshall and G. Lang, Proc. Phys. Soc. (London), 87, 3 (1966).
22. T. H. Moss, A. Ehrenberg, and A. J. Bearden, Biochemistry, 8, 4159 (1969).
23. P. Debrunner, in "Spectroscopic Approaches to Biomolecular Conformation," ed. D.
W. Urry, American Medical Association Press, Chicago, 1969.
24. C. L. Thompson et al., J. Biochem. 139, 97 (1974) and references therein.
25. R. H. Holm, Endeavour, 1975, 38 and references therein.
26. R. H. Holm et al., J. Amer. Chem. Soc., 96, 2644 (1974) and references therein.
27. G. M. Bancroft, Coord. Chem. Rev., 11, 247 (1973) and references therein.
28. G. M. Bancroft and K. D. Butler, Inorg. Chim. Acta, 15, 57 (1975).
29. A. P. Marks, R. S. Drago, R. H. Herber, and M. J. Potasek, Inorg, Chem., 15,259 (1976).
30. W. H. Morrison, Jr., and D. N. Hendrickson, Inorg. Chem., 13, 2279 (1974).
SERIES "M6ssbauer Effect Methodology," Volumes I to 9, Plenum Press, New York, 1965.
Exercises 647
1. What effect does increasing electron density at the nucleus have on the relative energies EXERCISES
of the ground and excited states of 57 Fe and 119 Sn? Explain the expected isomer shift
from this change in terms of effective nuclear charge radii of these states.
2. Suppose you read an article in which the author claimed that the two peaks in a MB
spectrum of a low spin iron(III) complex were the result of Jahn-Teller distortion.
Criticize this conclusion.
3. Draw the structure for SnF 4 and explain why quadrupole splitting is observed in this
compound but not in SnCl 4 .
4. Suppose you were interested in determining whether Sn-O or Sn-S 7T bonding were
present, and which was greater, in the compounds (CH 5 )3 SnOCH3 and
(C6HI) 3SnSCH 3 . Describe experiments involving MB spectroscopy that might shed
light on this problem.
5. Would you expect the AEQ value to be greatest in SnCl 2 Br 2 , SnCl3 Br, or SnF 3 I?
Explain.
6. The product obtained from the reaction of ferrous sulfate and potassium ferricyanide
gives rise to the spectrum below. Interpret this spectrum.
0 1.0
mm sec-1
7. Using a point charge expression, indicate the three diagonal components of the electric
field gradient for the two isomers of MA 3 B3 . Predict the sign of the quadrupole
splitting in the two isomers.
8. (CH 3)2 SnCl2 is a chloro-bridged polymer with octahedral coordination about tin. The
methyl groups are trans. With the CH 3 -Sn bond as the z field gradient axis, predict
the signs of q and e2 Qq.
9. Which componds would have the largest quadrupole splittings for the starred atom:
a. cis or trans *Fe(CO)4Cl2?
b. (CH 3)3 *SnCl (C3 ,) or polymeric, chloro-bridged (CH 3 )3 *SnCl (D 3, about tin)?
c. high spin cis or trans *Fe(NH3)4 C 2 ?
10. The M6ssbauer spectrum for Fe(CO), at liquid nitrogen temperature is shown below.
-4.0 0 4.0
1
mm sec-
a. Why are the unusual units on the x-axis equivalent to energy?
b. What gives rise to the doublet observed? Show (and label) the energy levels involved.
11. The enzyme putidaredoxin has sites with two iron atoms. The oxidized form gives
rise to the spectrum shown below:
0.000
0.005- *
0.011 -
0.016 - 0S
c 0.022 -
0.0270- 01
S0.033 -
< 0.038 -
0.044 -
0.049 -
0.055-
-0.1 0 +0.1
Velocity (cm/sec)
The g-tensors are anisotropic. Does the enzyme consist of a single type of iron site
or two different iron sites?
12. One of the M6ssbauer spectra below is K 3 Fe(CN),. The other is K4 Fe(CN),. Which
is which? Why?
150-
140-
130
-0.1 0 0.1 0.2 -0.1 0 0.1 0.2

Exercises 649
13. An article [J. Amer. Chem. Soc., 97,6714 (1975)] dealt with a series of square-pyramidal
iron(III) complexes having an N4 S (the sulfur donor is axial) set of donor atoms.
These complexes varied considerably in electronic structure, depending upon the exact
nature of the N 4 macrocyclic ligand and the axial ligand, as shown by various
spectroscopic measurements.
a. Using one-electron energy level diagrams, indicate which spin states might
reasonably be expected for square-pyramidal iron(III). Remember that the relative
energies of the d orbitals may vary slightly as a function of the donor set.
b. Using spectra from the figure below, assign probable ,spin states to complexes A, B,
and C. Explain what features of the M6ssbauer spectra influenced your decision,
and why.
(A)
(B)
-2 -1 0 1 2
Velocity (mm/sec)
16 Ionization Methods:
Mass Spectrometry,
Ion Cyclotron Resonance,
Photoelectron
Spectroscopy
Mass 16-1 INSTRUMENT OPERATION AND PRESENTATION OF

SPECTRA
Spectrometry
There are many different types of mass spectrometers available, and the details
of the construction and relative merits of the various types of instruments have
been described.(1 -6) Most of the basic principles of mass spectrometry can be
illustrated with the aid of Fig. 16-1. The sample, contained in a reservoir, is
added through the port, enters the ion source (a), and passes through the electron
beam at (c). (The beam is indicated by a dotted line.) Interaction of the sample
with an energetic electron produces a positive ion, which moves toward the
accelerating plates (d) and (e) because of a small-potential difference between the
back wall (inlet) and the front wall of this compartment. The back wall, which
To vacuum:
650
FIGURE 16-1 A schematic of a 1800 deflection mass spectrometer.
is positively charged relative to the front, attracts and discharges the negative
ions. The positive ions pass through (d) and (e), are accelerated by virtue of a
large potential difference (a few thousand volts) between these plates, and leave
the ion source through (b). Charged ions move in a circular path under
the influence of a magnetic field. The semicircle indicated by (f) is the path
traced by an accelerated ion moving in a magnetic field of strength H. The radius
of the semicircle, r, depends upon the following: (1) the accelerating potential, V
[i.e., the potential difference between plates (d) and (e)]; (2) the mass, m, of the
ion; (3) the charge, e, of the ion; and (4) the magnetic field strength, H. The
relationship between these quantities is given* by equation (16-1):
m H2 r2
- 2V(16-1)
e 2V
When the ions pass through slit (g) into the detector, a signal is recorded. The
source and the path through which the ions pass must be kept under 10' mm
of mercury pressure in order to provide a long mean free path for the ion. The
sample vapor pressure in the inlet should be at least 10-2 mm Hg at the
temperature of the inlet system, although special techniques permit use of
lower pressures. In general, only one or two micromoles of sample are required.
Since H, V, and r can be controlled experimentally, the ratio m/e can be
determined. Note that a dipositive ion of mass 54 gives rise to the same m/e as
a monopositive ion of mass 27. Under usual conditions for running a mass
spectrometer, most of the ions are produced as singly charged species. Doubly
charged ions are much less frequently encountered, whereas more highly charged
ions are not present in significant concentrations.
Obtaining the mass spectrum consists in determining the m/e ratio for all
fragments produced when a molecule is bombarded by a high intensity electron
beam. To do this, the detector slit could be moved and the value of r measured for
all particles continuously produced by electron bombardment in the ion source.
This is not feasible experimentally. It is much simpler to vary H or V continuously
[see equation (16-1)] so that all particles eventually travel in a semicircle of fixed
radius. The signal intensity, which is directly related to the number of ions striking
the detector, can be plotted as a function of H or V, whichever is being varied.
In practice it is easiest to measure a varying potential, so the field is often held
constant. If H and r are constant, V is inversely related to m/e [by equation
(16-1)] and m/e can be plotted versus the signal intensity to produce a conven-
tional mass spectrum. A narrow region of a typical spectrum is illustrated in Fig.
16-2.
On instruments in which the potential is varied, the results are readily
presented in terms of m/e values. The accuracy varies with mass, as can be seen
by comparing the potential difference corresponding to m/e values of 400
* This equation is simply derived: The potential energy of the ion, eV, is converted to kinetic
energy (1/2 )mV2 (where m is the mass and v the velocity). At full acceleration,
2
eV= 1mv (16-2)
In the magnetic field the centrifugal force mv 2/r is balanced by centripetal force Hev. Solving
Hev mV2/r for r yields r = mv/eH. Using this equation to eliminate v from (16-2) produces (16-1).
652 Chapter 16 Ionization Methods
FIGURE 16-2 Mass spectrum of

fragments with m/e in the range from
40 to 48. The peaks have been
automatically attenuated, and the
number of horizontal lines above
each peak indicates the attenuation
factor.
40 41 42 43 44 45 46 47
m/e
and 401 with the difference for m/e values of 20 and 21 [calculated by using
equation (16-1)]. With modern low-resolution instruments, one can count on
+0.4 to + 1 m/e accuracy up to 1000 mass units. For high-resolution work,
accuracy to 2 or 3 ppm can be obtained up to very high masses (>3000) if
reference compounds are available that have accurately known peaks within
10 to 12% of the unknown mass peaks. Perfluorokerosine is a common
reference compound, with many peaks that are accurately known up to about
900 mass units.
The mass spectrum obtained from the instrument may be a long and
cumbersome record. As a result, the data are often centroided and summarized
by a bar graph that plots intensity (relative abundance of each fragment) on
the ordinate and the m/e ratio on the abscissa. The relative abundance is given
as the percent intensity of a given peak relative to the most intense peak in
the spectrum. Often it is informative to express the intensity in terms of total
ionization units, S. This is an expression of the percentage that each peak
contributes to the total ionization, and is obtained by dividing the intensity of
a given peak by the sum of the intensities of all the peaks. When the spectrum
has not been obtained over the range from 1 to the molecular weight of the
material, the lower limit is indicated on the ordinate as a subscript to E (e.g.,
Y-, means that the spectrum was recorded from 12 to the molecular weight).
A typical graph is illustrated in Fig. 16-3. For the comparison of the intensity
of various peaks in the same spectrum, the relative abundance is adequate. For
comparison of the intensity of peaks in different spectra, total ionization units
should be employed. In general, a peak with an intensity equal to 0.5% of Y
is easily detected.
Most modern spectrometers contain computers to store and process the
large amount of information from molecular fragmentation and to control the
experimental variables involved in running the instrument. Accurate masses can
be obtained with computer interfacing by introducing standards with the
sample. The computer recognizes the standard peaks and employs them to
determine the mass of the unknown peaks. Computer libraries 7 ) are available
to match the sample spectra with that of known compounds in 70-eV electron
impact ionization.
100 38.2
-30
FIGURE 16-3
50- -20 .E Representation of the mass spectrum
of ethanol.
-10
10 20 30 40 50
m/e
Our next concern is with the method of indicating the resolving power of
the instrument. For many inorganic and organic applications, it is necessary
to know the m/e ratio to an accuracy of one unit (i.e., whether it is 249 or 250).
The resolution of the instrument is sometimes expressed as m/Am, where two
peaks m and m + Am are separately resolved and the minimum intensity
between the two peaks is only 10% of the total of m. For example, a resolution
of 250 means that two peaks with m/e values of 251 and 250 are separated and
that at the minimum between them the pen returns to the base line to within
10% of the total ion current (plotted as intensity in Fig. 16-2). Instruments with
poor resolution will not do this with high mass peaks, and the magnitude of
the m/e value for which peaks are resolved is the criterion for resolution.
Modern mass spectrometers provide a wide range of options in inlet
systems, ion sources, mass analyzers, detectors, and signal processors. Batch
inlet systems volatilize a sample externally and allow it to leak into the
ionization region. The sample must have an appreciable vapor pressure below
~500 C. The direct probe inlet introduces non-volatile and thermally unstable
compounds with a sample probe that is inserted through a vacuum lock. The
output from a chromatographic column can also be used as the inlet in GC-MS
systems.
Most commercial mass spectrometers are equipped with accessories that
allow use of several ion sources interchangeably. Some common ones are listed
in Table 16-1. The desorption types use a sample probe from which energy
directly transfers ions or molecules in the condensed phase to ions in the gas
phase. Non-volatile and thermally fragile molecules can be studied. Electron
impact, EI, uses a beam of energetic electrons to ionize the gaseous molecule.
Field ionization and field desorption employ large voltages (10-20 kV) to
produce large electric fields (10' V/cm) to ionize molecules. Gaseous samples
are used in the former and non-volatile samples on a probe in the latter.
Fragmentation of the molecule is minimized with the latter technique and the
molecular ion, M', and one with a mass one unit larger, (M + 1)*, are the
major peaks in the spectrum. In chemical ionization, CI, the sample is bom-
barded by positive ions produced by electron bombardment of a gaseous atom
or molecule (often CH 4 ) present with the sample and in 1000- to 10,000-fold
TABLE 16-1. Some Common Ion Sources

Name Abbreviation Type Ionizing Agent
electron ionization El gas phase energetic electrons

field ionization FI gas phase high-potential electrode
chemical ionization CI gas phase reagent positive ions or
electron capture
fast atom bombardment FAB desorption" energetic atoms
field desorption FD desorption' high-potential electrode
laser desorption LD desportion laser beam
plasma desorption PD desorption' high-energy fission
fragments from 252Cf
secondary ion mass SIMS desorption 1 to 20 keV ions
spectrometry
thermal desorption TIMS desorption' heat
Samples as solids or solutions

Samples as solids, gases, or solutions.
excess. We can illustrate the reactions in chemical ionization that lead to

product ions with CH 4 as an example. The CH 4 + ion formed by ionization
reacts with another methane molecule forming CH,*. The CH,' or CH 4 can
react with substrate SH forming SH 2 + and CH 4 or CH 3 forming the substrate
(M + 1)* ion. With some substrates, CH,* or CH 4 can abstract hydride
forming H 2 and S*, where S' is the (M - 1)' ion. In both instances, reactions
of the ion with the sample lead to sample ions. Cluster ions can also form in
CI, e.g., reactions occur to form (M + C2 H,)* and (M + C3 H,)* peaks in
methane CI. Less fragmentation results with CI than with EL. In fast atom
bombardment (FAB), the sample in a condensed state (often a glycerol matrix)
is ionized by bombardment with energetic Xe or Ar atoms. Molecular ions are
often formed even for large molecular-weight samples.
Analyzers are commonly radio-frequency filters or electromagnets in
combination with electrostatic fields. The radial magnetic field shown in Fig.
16-1 employs single-focusing magnetic sector analysis. Double focusing analy-
zers first pass the beam through a radial electrostatic field and then the radial
magnetic field. The electrostatic field focuses particles with the same kinetic
energy on a second slit which introduces the beam into the magnetic field.
Much better resolution is achieved. The quadrupole analyzer consists of four,
short, parallel metal rods arranged symmetrically around the beam. One pair
of opposite rods is attached to the positive terminal of a variable dc source and
the other pair to the negative terminal. On top of this, variable radio-frequency
ac potentials, that are 180 out of phase are applied to each pair of rods. The
combined fields cause the charged particles in the beam to oscillate about a
central axis describing their direction of travel. Only fields with a certain mass
to charge ratio pass through the rods. The others collide with one of the rods.
The m/e ratio is scanned by varying the frequency of the ac source or by varying
the potential of the two sources keeping their ratio and frequency constant.
The ion trap uses a field generated by a sandwich geometry with a ring electrode
in the middle and caps on each end to trap and store ions of a selected
mass-to-charge ratio. The mass spectrum is produced by varying the electric
field to sequentially eject ions of different m/e ratios for detection.
16-2 Processes that can Occur When a Molecule and a High Energy Electron Combine 655
In a time-of-flight instrument, the semicircular path and magnet are

replaced by a straight path (no magnetic field). The ions are not produced
continuously but in spurts and are allowed to diffuse toward the detector. Heavy
ions will move more slowly than light ions. The spectrum is a plot of intensity
versus the time of flight of the particle. The application of the Fourier transform
technique to mass spectrometry has many of the advantages of applying this
technique to other spectroscopies. This method will be discussed in the next
section on ion cyclotron resonance.
16-2 PROCESSES THAT CAN OCCUR WHEN A

MOLECULE AND A HIGH ENERGY ELECTRON
COMBINE
It is important to emphasize the rather obvious point that a detector analyzes

only those species impinging on it. We must, therefore, be concerned not only
with the species produced in the ionization process but also with the reactions
these species may undergo in the 10 - I sec required to travel from the accelerating
plates to the detector. When a molecule, A, is bombarded with electrons of
moderate energy, the initial processes that can occur are summarized by equa-
tions (16-3) to (16-5):
A + e- -+ A + 2e- (16-3)
A + e - A"+ + (n + 1)e (16-4)
A + e - A- (16-5)
The process represented by reaction (16-3) is the most common and the most
important in mass spectrometry. It will occur if the energy of the bombarding
electron is equal to or higher than the ionization energy of the molecule (7 to
15 eV). When the energy of the bombarding electron beam is just equal to the
ionization potential, all of the electron's energy must be transferred to the
molecule to remove an electron. The probability of this happening is low. As the
energy of the bombarding electron increases, the probability that a collision will
induce ionization increases, and a higher intensity peak results. As the energy of
the bombarding electron is increased further, much of this excess energy can be
given to the molecular ion that is formed. This excess energy can be high enough
to break bonds in the ion, and fragmentation of the particle results. The
acceleration potential of the bombarding electron that is just great enough to
initiate fragmentation is referred to as the appearancepotential of the fragment
ion. When the electron energy is large enough, more than one bond in the
molecule can be broken. The following sequence summarizes the processes that
can occur when a hypothetical molecule B-C-D-E is bombarded with an
electron:
(a) Ionization process
BCDE + e- --* BCDE* + 2e- (16-6)

'I
(b) Fragmentation of the positive ion
BCDE* - B+ + CDE- (16-7)
BCDE* -> BC* + DE- (16-8)
BC+ -* B* + C- (16-9)
or BC+ - B- + C+ (16-10)
BCDE*- DE ++BC (16-11)
DE+ * D- + E (16-12)
or DE*-*D-i+E* (16-13)
BCDE* ->BE+ + CD- etc. (16-14)
BCDE* -- CD+ + BE- etc. (16-15)
(c) Pair production
BCDE + e --* BC* + DE- +e~ (16-16)
(d) Resonance capture
BCDE + e --> BCDE- (16-17)
Other modes of cleavage are possible, but only positively charged species will
travel to the detector and give rise to peaks in the mass spectrum. For the scheme
above, peaks corresponding to B*, BC+, C+, DE*, D+, E+, BE+, and CD+
will occur in the spectrum if B, C, D, and E have different masses. More energy
will have to be imparted to the BCDE* ion to get cleavage into B+, C-, and
DE- [equation (16-9)] than for cleavage into BC* + DE- [equation (16-8)].
In equation (16-14), the ion has rearranged in the dissociation process,
leading to fragments that contain bonds that are not originally present in the
molecule BCDE. These rearrangement processes complicate the interpretation
of a mass spectrum, and experience is necessary to be able to predict when
rearrangements will occur e.g., alkenes commonly rearrange. In general, re-
arrangements are invoked to explain the occurrence of peaks of unexpected mass
or unexpected intensity.
Ion-molecule reactions can give rise to mass peaks in the spectrum that are
greater than the molecular weight of the sample. This process is represented by
equation (16-18):
BCDE* + CDE- -* BCDEC+ + DE (16-18)
The reaction involves a collision of the molecular ion with a neutral molecule
and as such is a second-order rate process, the rate of which is proportional to
the product of the concentrations of the reactants. The intensity of peaks resulting
16-3 Fingerprint Application 657
from this process will depend upon the product of the partial pressures of BCDE +
and CDE. Examination of the spectrum at different pressures causes variation
in the relative intensities of these peaks, and as a result, the occurrence of this
process can be easily detected.
The reactions discussed above are all unimolecular decay reactions. Produc-
tion of ions by electron bombardment often involves loss of the least tightly held
electron, and ions are often formed in vibrationally excited states that have an
excess of internal energy. In some molecules of the sample, a low energy electron
is removed, leaving an ion in an excited electronic state. The excited state ion
can undergo internal conversion of energy, producing the electronic ground state
of the ion having an excess of vibrational energy. The molecule could dissociate
in any of the excited states involved in the internal conversions associated with
the radiationless transfer of energy. In other molecules of the sample, ions are
formed with energy in excess of the dissociation energy. In this case, the ion will
fragment as soon as it starts to vibrate. Thus, in a given sample, ions with a wide
distribution of energies are produced and many mechanisms are available for
fragmentation processes. It is informative to compare the time scales for some
of the processes we have been discussing. The time for a bond vibration is
~10- " sec, the maximum lifetime of an excited state ion is - 10 -8 sec, and the
time an ion spends in the mass spectrometer ion chamber is 10-' to 10-6 sec.
There is ample time for the excess electronic energy in an ion to be converted
into an excess of vibrational energy in a lower electronically excited state.
Accordingly, we view the processes in the ionization chamber as producing
molecular ions in different energy states, which undergo rapid internal energy
conversion to produce individual ions with varying amounts of excess energy.
Fragmentation takes place via a first order process at different rates, depending
on the electronic state and excess vibrational energy of the individual ion. This
is why all of the different processes represented above for the fragmentation of
the ion BCDE+ can occur and be reflected in the mass spectrum.
16-3 FINGERPRINT APPLICATION
The fingerprint application is immediately obvious. For this purpose an electron

beam of 70 eV is usually employed to yield reproducible spectra, for this
accelerating potential is above the appearance potential of most fragments. As
indicated by equations (16-6) to (16-16), a large number of different fragmenta-
tion processes can occur, resulting in a large number of peaks in the spectra of
simple molecules. Figure 16-3 contains the peaks with appreciable intensity that
are found in the mass spectrum of ethanol. Counting the very weak peaks, which
are not illustrated, a total of about 30 peaks are found. These weak peaks are
valuable for a fingerprint application but generally are not accounted for in the
interpretation of the spectrum (i.e., assigning the fragmentation processes leading
to the peaks).
For a fingerprint application, the more peaks the better. Accordingly, El
methods are preferred. Spectral libraries for El at 70 eV are available in text""
17
and computer formo for matching an unknown spectrum to a known one. The
relative intensities are of less help in unknown identification for they vary with
instrument and with experimental conditions. It is best to make a direct
comparison of the spectrum of the unknown compound with that of a known

sample on the same instrument if a decision is to be based on relative intensities.
An interesting fingerprint application is illustrated in Fig. 16-4, where the
mass spectra of the three isomers of ethylpyridine are indicated. Pronounced
differences occur in the spectra of these similar compounds that are of value in
the fingerprint application. On the other hand, similarities in the spectra of
hydrocarbons makes the fingerprint application of limited utility for this class
of compounds. Optical isomers and racemates give rise to identical spectra.
Impurities create a problem in fingerprint applications because the major
fragments of these impurities give rise to several low intensity peaks in the
N CH2-CH 3
Mol. wt. 107
FIGURE 16-4 Mass spectra

of three isomers of CH
2-CH,
ethylpyridine. [Copyright @
1960 McGraw-Hill Book
Company. From K. Biemann,
"Mass Spectrometry."
Reproduced by permission.]
(C)
50t
460 /e0 00
m/e
16-4 Interpretation of Mass Spectra 659
spectrum. If the same material is prepared in two different solvents, the spectra
may appear to be quite different if all solvent has not been removed. Contamina-
tion from hydrocarbon grease also gives rise to many lines.
16-4 INTERPRETATION OF MASS SPECTRA
The interpretation of a mass spectrum involves assigning each of the major peaks
in a spectrum to a particular fragment. An intense peak corresponds to a high
probability for the formation of this ion in the fragmentation process. In the
absence of rearrangement [equation (16-14)], the arrangement of atoms in the
molecule can often be deduced from the masses of the fragments that are
produced. For example, a strong peak at m/e = 30 for the compound methyl
hydroxylamine would favor the structure CH 3 NHOH over H2 NOCH 3 because
an m/e = 30 peak could result from cleavage of the O-N bond in the former
case but cannot result by any simple cleavage mechanism from the latter
compound. The higher mass fragments are usually more important than the
smaller ones for structure determination.
It is often helpful in assigning the peaks in a spectrum to be able to predict
probable fragmentation products for various molecular structures. The energy
required to produce a fragment from the molecular ion depends upon the
activation energy for bond cleavage, which is often related to the strength of the
bond to be broken. The distribution of ions detected depends not only on this
but also on the stability of the resulting positive ion. In most cases it is found
that the stability of the positive ions is of greatest importance. This stability is
related to the effectiveness with which the resulting fragment can delocalize the
positive charge. Fragmentation of HOCH2 CH 2 NH 2 + can occur to produce
-CH 2 0H and CH 2NH 2 + (m/e = 30) or -CH 2NH 2 and CH 2 OH+ (m/e = 31).
Since nitrogen is not as electronegative as oxygen, the resonance form CH 2 =
NH 2 + contributes more to the stability of this ion than a similar form, CH 2 =
0H, does to its ion. As a result, charge is more effectively delocalized in the
species CH 2 NH 2 ' than in CH 2 OH+, and the mle = 30 peak is about ten times
more intense than the m/e = 31 peak. Charge is not stabilized as effectively by
sulfur as it is by oxygen because carbon-sulfur 7r bonding is not as effective as
carbon-oxygen 7r bonding. Thus, the m/e = 31 peak for CH 2 OH+ from
HSCH 2 CH 2OH has about twice the intensity of the m/e = 47 peak that arises
from CH 2SH+.
Rearrangements of the positive ion will occur when a more stable species
results. For example, the ion CH 2 rearranges to + and
the benzyl cation rearranges to the tropylium ion, C7 H7 . An intense m/e = 93

peak from this fragment is seen in alkyl benzenes.
The production of many different fragments is often helpful in putting
together the structure of the molecule. However, one must employ caution even
in this application. The ion produced in the ion chamber undergoes many
vibrations, during which rearrangement could occur to produce bonds that did
not exist in the parent compound [see, e.g., equation (16-14)]. The production
of all these different ions makes it difficult to determine the chemical processes
that lead to the various peaks. This in turn makes it difficult to infer the influence
that bond strength or other properties of the molecule have on the relative
abundances of the ion fragments formed. A quantitative treatment of mass
spectrometric fragmentation has been attempted and is referred to as the
quasi-equilibrium theory." ) The internal energy is distributed over all the
available oscillators and rotators in the molecule, and the rates of decomposition
via different paths are calculated. A weighting factor or frequency factor (i.e., an
entropic term) is given to each vibration level. The full analysis is complex for
a molecule of reasonable size. Approximations have been introduced, leading to
a highly parametrized approach."'")
The mass spectrometric shift rule"') has been of considerable utility in
elucidation of the structure of alkaloids and illustrates a basic idea of general
utility. If there are low energy pathways for the breakdown of a complex molecule
and this breakdown is not influenced by the addition of a substituent, the location
of the substituent can often be determined. This is accomplished by finding an
increase in the molecular weight of the fragment to which the substituent is
bonded that corresponds to the weight of the substituent or a characteristic
fragment of the substituent.
Mass spectrometry is used for routine sequencing of small peptides. The
interested reader is referred to the review described in reference 12.
The low volatility of many substances hampers their analysis by mass
spectroscopy. The volatility can often be increased by making derivatives of the
polar groups in the molecule; e.g., carboxyl groups can be converted to methyl
esters or trimethylsilylesters. Field ionization techniques (vide infra) are also
advantageous for this problem.
The combination of mass spectroscopy with GLC provides an excellent
method for analysis of mixtures. Very small amounts of material are needed. The
mass spectrometer may be used as the GLC detector, and numerous mass spectra
can be accumulated as each component emerges from the column. A partially
resolved GLC peak is readily detected by the change in mass spectra of the peak
with time.
Many more examples and a thorough discussion of factors leading to stable
ions produced from organic compounds are contained in the textbook references
at the end of the chapter and reference 9. Generalizations for predicting when
rearrangements are expected are also discussed. If, starting with a given structure,
one can account for the principal fragments and assign the peaks in the mass
spectrum by invoking a reasonable fragmentation pattern, this assignment
amounts to considerable support for that structure.
16-5 EFFECT OF ISOTOPES ON THE APPEARANCE OF A

MASS SPECTRUM
When the spectrum of a compound containing an element that has more than
one stable, abundant isotope is examined, more than one peak will be found for
each fragment containing this element. In the spectrum of CH 3 Br, two peaks of
nearly equal intensity will occur at m/e values of 94 and 96, corresponding mainly
79
to (CH 37 9Br)* and (CH 3 81 Br)+. The abundances of Br and "Br are almost
the same (50.54 versus 49.46%), so two peaks of nearly equal intensity, separated
by two mass units, will occur for all bromine-containing fragments. In a fragment
containing two equivalent bromine atoms, a triplet with ratios 1:2:1 would result
16-5 Effect of Isotopes on the Appearance of a Mass Spectrum 661
from different combinations of isotopes. In addition to these peaks, there will be

small peaks resulting from the small natural abundances of D and "C, corre-
sponding to all combinations of masses of "C, "C, D, H, 79 Br, and 8 'Br. The
resulting cluster of peaks for a given fragment is important in establishing the
assignment of the peaks to a fragment. Their relative intensities will depend upon
the relative abundances of the various naturally occurring isotopes of the atoms
in the fragment; e.g., CO' can consist of mass fragments at 28, 29, 30, and 31.
The relative abundances of these fragments can be calculated from simple
probability theory. ( 3" 4 ) Computer programs have been reported to carry out
these calculations.(15, 16 ) These characteristic patterns are quite useful in assigning
spectra of molecules that contain an atom with more than one abundant isotope.
Molecules containing transition metals often give such isotope patterns. Use of
the 3 C peaks enables one to determine the number of carbon atoms in a
fragment.
The advantages of high resolution mass spectrometry can be illustrated"?)
by the incorrect assignment of a peak at 56 in the low resolution spectrum of
Fe[(C 2 H5 ) 2 NCS 2 ] 3 to iron. At high resolution, a peak is expected at 55.9500,
but none is found. Instead, one is obtained at 56.0350, which is assigned to the
fragment C3HN.
Another important application of the mass spectrometer involving isotopes
is the study of exchange reactions involving nonradioactive isotopes. The product
of the exchange from labeled starting material is examined for isotope content
as a function of time to obtain the rate of exchange. The product or starting
material can be degraded to a gaseous material containing the label, and the
isotopic ratio is obtained from the mass spectrum. These materials may also be
examined directly, and the location and amount of label incorporated can be
deduced from an analysis of the change in spectrum of various fragments. By
determining which peaks in the spectrum change on incorporation of the isotope,
one can determine which parts of a molecule have undergone exchange. In the
reaction of methanol with benzoic acid, it has been shown by a tracer study
involving mass spectral analysis that the ester oxygen in the product comes from
methanol:
0 0
18 // 15
CHC-OH + CH 3 18 0H - CHC-' 0CH3 + H 2 0
In another interesting application it was shown that the following exchange

reactions occurred:
BF 3 + BX 3 - BX 2 F + BF 2 X
3RBX 2 + 2BF 3 - 3RBF 2 + 2BX 3 (also BFnXm)
3R 2 BX + BF 3 - 3R2 BF + BX 3 (also BFX.)
where R is alkyl or vinyl and X is Cl or Br. Fragments corresponding to

the products were obtained, -although only starting materials were re-
covered on attempted separation.( 18) A four-center intermediate of the type
R Cl F
B B was proposed for the exchange. In order to determine
Cl F F
whether or not alkyl groups were exchanged in the reaction:
3RBX 2 + 2BF 3 -* 3RBF 2 + 2BX 3 (and BX, F,)

0
the boron trifluoride was enriched in "0B. The absence of enrichment of " B in
fragments in the mass spectrum containing alkyl or vinyl groups enabled the
authors to conclude that neither alkyl nor vinyl groups were exchanged under
conditions where RBX 2 species were stable.
It should be pointed out that in all of the above applications it is not
necessary to label the compound completely. A slight enrichment will suffice.
16-6 MOLECULAR WEIGHT DETERMINATIONS;

FIELD IONIZATION TECHNIQUES
Determination of the molecular weight requires methods that will produce the
parent ion peak. The dramatic difference in El and FL methods is illustrated with
the mass spectra in Fig. 16-5. The M + 1 peak occurs from addition of a proton
in the presence of the reagent ion. In some cases, proton abstraction leads to an
M - 1 peak. 9
In Fig. 16-6, the spectra of glutamic acid'1 a)
HOOCCH(NH 2 )CH 2 CH 2 COOH
80
60
-S 40
M =150
0
HOH2C H rel. int. = 0.2%
20
H 86
FIGURE 16-5 (A) Formula 132
1 119 1 k -I- _ _
of D-ribose. (B) Electron OH OH I fI I I I I I I I I I I 1
bombardment mass 20 40 60 80 100 120 140 160
MW = 150 mle
spectrum. (C) Field ionization
spectrum. (A) (B)
29
43
61 74 133
20 40 60 80 100 120 140 160

m/e-
(C)
16-7 Evaluation of Heats of Sublimation and Species in the Vapor Over High Melting Solids 663
(M - H 2 0 + H)+
M + H)*
100- 130 100 148
Q>
c80-
8 80-
C 60 C6
:3 60 -
84
>40 + +H)+ 40-
20 COOH * 148 2
-7 0 45 20
Z0
0 40 80 120 ^160 0 40 80 120 160

m/e m/e
(A) (B)
FIGURE 16-6 Mass spectra for glutamic acid: (A) field ionization; (B) field desorption.
obtained by field ionization and field desorption methods are compared. Very
little fragmentation occurs with field desorption. Either method is adequate
for molecular weight determination. FAB is the most frequently used method
for molecular weight determination of non-volatile materials giving dominant
(M + H)+ and (M - H)* ions for organic materials. With organic materials and
organometallic compounds, significant fragment ion information is obtained
providing structural information. A comparison of the results on organometallic
compounds from various ionization techniques is reported.(19 )
16-7 EVALUATION OF HEATS OF SUBLIMATION

AND SPECIES IN THE VAPOR OVER HIGH
MELTING SOLIDS
Evaluation of the heat of sublimation is based upon the fact that the intensity
of the peaks in a spectrum is directly proportional to the pressure of the sample
in the ion source. The sample is placed in a reservoir containing a very small
pinhole (a Knudsen cell), which is connected to the ion source so that the only
way that the sample can enter the source is by diffusion through the hole. If the
cell is thermostated and enough sample is placed in the cell so that the solid
phase is always present, the heat of sublimation of the solid can be obtained by
studying the change in peak intensity (which is related to vapor pressure) as a
function of sample temperature. The small amount of sample diffusing into the
ion beam does not radically affect the equilibrium in the cell. Some interesting
results concerning the nature of the species present in the vapor over some high
melting solids have been obtained from this type of study. Monomers, dimers,
and trimers were found over lithium chloride, while monomers and dimers were
found in the vapor over sodium, potassium, and cesium chloride.(4'
The species Cr, CrO, CrO 2 , 0, and 02 were found over solid Cr 2 O3 .
Appearance potentials and bond dissociation energies of these species are
reported. 2" The vapors over MoO 3 were found to consist of trimer, tetramer,
and pentamer. Vapor pressures, free energy changes, and enthalpies of sublima-
tion were evaluated?.
16-8 APPEARANCE POTENTIALS AND IONIZATION

POTENTIALS
As mentioned earlier, the molecular ion is produced whenever collision occurs

with an electron with energy equal to or greater than the ionization energy of
the molecule. A typical curve relating electron energy to the number of ion
fragments of a particular type produced (i.e., relative intensity of a given peak)
is illustrated in Fig. 16-6. This is referred to as an ionization efficiency curve. At
electron energies well below the ionization energy, no ions are produced. When
the energy of the electron beam equals the ionization energy, a very low intensity
peak results, for in the collision all of the energy of the electron will have to be
imparted to the molecule, and this is not too probable. As the electron energy
is increased, the probability that the electron will impart enough energy to the
molecule to cause ionization is increased, and a more intense peak results until
a plateau finally occurs in the curve. The tail of this curve at low energies results
because of the variation in the energies of the electrons in the bombarding beam.
Therefore, the curve has to be extrapolated (dotted line in Fig. 16-7) to produce
the ionization energy. Various procedures for extrapolation and the error
introduced by these procedures have been discussed in detail.(") When the peak
observed is that of the molecular ion, e- + RX -> RX+ + 2e-, the ionization
energy of the molecule can be obtained by extrapolation of the ionization
efficiency curve. When the peak is that of a fragment, extrapolation of the
ionization efficiency curve produces the appearance potential of that fragment.
For example, if the peak being investigated is that of the fragment R+ from the
molecule R-X, the appearance potential, AR+, is obtained by extrapolation of
the ion efficiency curve for this peak. The appearance potential is related to the
following quantities:
AR+ = DR x + IR + Ek + Ee (16-19)
where DR-x is the gas phase dissociation energy of the bond R-X; 'R is the
ionization potential of R; Ek is the kinetic energy of the particles produced; and
E, is the excitation energy of the fragments (i.e., the electronic, vibrational and
t
C
FIGURE 16-7 An ionization i

efficiency curve.
Inc. electron energy --

16-9 The Fourier Transform Ion Cyclotron Resonance Technique
rotational energy if the fragments are produced in excited states). Generally, E,

and E, are small and equation (16-19) is adequately approximated by:
AR. = DR x + IR (16-20)
If DR-xis known, IR can be calculated from appearance potential data. Often,

IR is known, and DR -x can be calculated. The value for 'R must be less than Ix
for equation (16-20) to apply; otherwise, X* is dissociated or electronically
excited. Experiments of this sort provide one of the best methods for evaluating
bond dissociation energies but give less exact ionization energy data than can
be obtained by other means.
An article on the mass spectrometric study of phosphine and diphosphine
contains a nice summary of some of the information that can be obtained from
these studies. Ionization energies and appearance potentials of the principal
positive ions formed are reported. The energetics of the fragmentation processes
are discussed and a mechanism is proposed.("'
16-9 THE FOURIER TRANSFORM ION CYCLOTRON FTICR/MS

RESONANCE TECHNIQUE
In FTICR/MS, an ion is generated with mass, m, and charge, e, as previously

discussed for mass spectroscopy. In a uniform magnetic field, the magnetic force
acting on a particle with initial velocity causes the ion to follow a helical path
along the axis of the magnetic field. The frequency of the circular motion in
Hertz, i.e., the cyclotron frequency is given by
v, = eH/27Em (16-21)
The cyclotron frequency is independent of the velocity of the ion, v. The velocity
distribution will give rise to a distribution of helical circular radii, r, for ions
with the same cyclotron frequency according to
v, = v/r
For magnetic fields of 1 T, the cyclotron frequency falls in the r.f. range (0.01 to
7.00 MHz).
In an ICR trapping cell, two trapping plates perpendicular to the magnetic
field are used to prevent the ions from travelling along the field direction. The
plates are maintained at a potential of about + 1 V for positive ions and -1 V
for negative ions. The magnetic field and trapping potential constrain the ions
to a region of space in the center of the trap. Two parallel receiving plates and
two parallel transmitter plates lying along the axis of the field and in between
the trapping plates form a box shaped cell of six plates. The cell is mounted in
a high vacuum in a strong magnetic field (typically ;>1 tesla).
Application of an external oscillating field with a frequency corresponding
to the cyclotron frequency of an ion, ve, across the transmitter plates causes ions
of that mass [equation (16-21)] to move into resonance with the applied field
and spiral out to orbits of larger radii. The kinetic energy of the particle is
increased and its radius also increases according to
Ek = 2nr2 mver 2
(16-22)
As the resonance condition is fulfilled, the random distribution of phases for the
ions is changed to a "packet" of ions that all move simultaneously with the
applied field. If the applied field is turned off or its frequency changed so it moves
out of resonance with ve, the ion packet persists long enough to induce an image
current in the detector plate. The signal decays with time as collisions of the ion
with neutral molecules restore the original random distribution of phases. The
signal in the time domain contains information about the frequency of the ion
and the ion concentration. 2 3 )
A mass spectrum of the total ion population in the cell is detected by
applying a field to the transmitter plates whose radio frequency changes rapidly
over the range corresponding to the masses of interest [recall equation (16-21)].
As each ion undergoes resonance, a superposition of image currents is observed
in the detector circuit and stored in a computer. The transmit-detect cycle is
repeated many times and signal averaged. A Fourier transform of this data
produces a plot of amplitude versus frequency or mass [equation (16-21)] of all
the ions present, i.e., the mass spectrum. The mass limit of detection depends on
the magnetic field. At 3 T, good mass resolution up to 3000 amu results.
After the ions are formed in the cell and before detection, any one ion can
be excited or, if they absorb sufficient energy, be caused to spiral to such large
orbits that they are ejected from the cell. The excited ions collide with neutral
molecules and fragment, a phenomenon called chemically induced dissociation
(CID). This technique is advantageous for structure determination or for provi-
ding endothermicities of bond dissociative pathways. If the ejected ion is reformed
with time by a chemical reaction, the rate of the reaction can be studied. By
monitoring changes that occur in ion concentrations when a reactive species is
introduced into the trapped ion cell, reaction rates and equilibria for gas phase
reactions can be measured.123 24 ,
Adiabatic electron affinities for a metal complex, M-, can be studied by
measuring the equilibrium constant for the reaction:
M -+SO2 SO2 + M
and knowing that the EA of SO 2 is 1.097 eV. If the reaction does not occur other
reference compounds besides SO2 can be employed. If the reaction goes to
completion with a compound X and does not occur with SO2 , the electron affinity
of M- is bracketed between X and SO 2 . About 100 organic compounds have
been bracketed between known systems over a range of 3 eV by this technique."
Adiabatic ionization energies for a substance M are obtained in an analogous
fashion by studying the reaction
M+ + X T M + X*
where the IE of X is known.

Ligand exchange reactions have also been studied enabling one to obtain
26
ligand basicity orders toward various cations.( a)
ML+ + L' -+ ML'* + L
In some instances the equilibrium constant can be measured, and in other

instances the relative basicities are estimated by laddering. Several excellent
reviews provide details concerning the above applications.( 2 6) Analytical applica-
tions of ICR have also been reviewed.(2 7 )
16-10 INTRODUCTION Surface

In this section we include techniques that, when applied to solids, provide Science
information primarily about the surface of the material. Some of the methods Techniques
can also be used to study gaseous and liquid samples. UPS is included in this
section even though it is used primarily on gaseous samples because of its
similarity to XPS. There is an immense literature background in this area. A
brief introduction to selected methods will be presented that discusses the
principles of operation and provides an illustration of the kind of information
obtained. Most of the methods use an ion, electron, or x-ray source and monitor
the energy or spatial resolution of the scattered (or resulting) ion, electron, or
photon beam produced after the source impinges on the sample. Since the source
can be very high in energy, induced reactions unfortunately can lead to the
detection of materials that are not characteristic of the sample. One must also
be concerned with the homogeneity of the surface when methods that produce
information about a small area are employed. Methodology that permits exami-
nation of the surface lateral and transverse planes addresses this concern. Finally,
one must be careful not to infer properties about the bulk sample from a surface
measurement unless it is known that the surface actually represents the bulk
property.
16-11 PHOTOELECTRON SPECTROSCOPY
Photoelectron spectroscopy has its basis in the photoelectric effect. There is a

threshold frequency, v, required in order for radiation impinging on a solid to
be able to eject an electron from the solid into a vacuum. The work function 4
is defined as the corresponding minimum energy required,
e4 = hv, (16-23)
where e is the charge of the electron and h is Planck's constant. At higher

frequencies of incident radiation, the emitted electrons have a kinetic energy
corresponding to this excess frequency, i.e.,
Eki = hv - e4 (16-24)
The emitted electrons with maximum kinetic energy come from the conduc-
tion band. Electrons in orbitals below the conduction band are bound and emerge
with less energy than those in the conduction band.
Ekin = hv - e# - EB (16-25)
where EB is the binding energy. Using photons whose energies are greater
than the work function and measuring the kinetic energy of the ionized electrons
produces the photoelectron spectrum. The spectrum is a plot of the number of
electrons emitted versus the kinetic energy of the electron, (see Fig. 16-8). The
peaks in the spectrum provide EB for various electron states of the surface species.
If a sample contains the same element in different chemical environments
(e.g., the nitrogen in N-N-O), different values for EB will result. Only a single
electron is ionized from a given molecule so the different signals arise from
different molecules. The difference in energy for the different EB values corre-
sponding to different environments is called the chemical shift.
The source photon is monochromatic. Depending upon whether x-ray or
ultraviolet radiation is employed, two different experiments result. The x-ray
source( 2 8a) experiment is called either x-ray photoelectron spectroscopy (XPS) or
28
electron spectroscopy for chemical analysis (ESCA). The ultraviolet source( b)
is called ultraviolet photoelectron spectroscopy (UPS or PES). XPS is concerned
with both valence and non-valence shell (core) electrons and UPS with valence
shell electrons. We shall discuss the applications of these two spectroscopic
methods separately.
XPS
Solids (including frozen solutions), gases, liquids, and solutions have been studied
by XPS. Binding energies, EB, of both core and valence electrons can be measured
as seen in equation (16-25). In the XPS experiment, the binding energies are
expressed relative to a reference level. Since the solid sample is in electrical contact
with the spectrometer, the Fermi level, e4, of the sample and spectrometer are
the same. As a result, only the spectrometer work function is needed to calculate
EB, referred to the Fermi level, from a measured kinetic energy. With an
insulating sample a buildup of positive charge can occur near the surface resulting
in a dipole layer and perturbation of the binding energies. An external standard
with a known core energy, (e.g., Au) can be deposited on the surface and binding
energies measured relative to the standard. The energies can also be measured
relative to the edge of the valence band.<2 9)
The binding energies of atoms vary with oxidation state and partial charges
on the atoms. Compilations exist(") that enable one to use binding energies to
determine these quantities by a fingerprint type of application. Generally, the
binding energies of cations increase with increasing oxidation state. Opposite
behavior is observed for PbO in comparison to PbO 2 and this is attributed to
differences in the Madelung potential in the two structures. Very small differences
in binding energy are found for Ag 2O compared to Ag and for Cu 2O compared
to Cu. On the other hand, the EB (Is) for Cu and CuO differ by 4.4 eV.
It is interesting to note that linear correlations have been found between
core-electron binding energies and M6ssbauer isomer shifts for compounds of
tin and iron.( 3 la)Also, a correlation of chlorine core binding energies with nuclear
31
quadrupole resonance frequencies has been found.( 1)
Attempts have been made to correlate the binding energies to orbital
energies from molecular orbital calculations. The basis for this approach is
Koopmans' theorem, which states that the vertical ionization energy for removal
of an electron from a molecular orbital corresponds to the negative of the
eigenvalue obtained from the molecular orbital calculation. It is thus assumed
that the molecular orbitals for the parent molecule and the ionized molecule are
the same. Any electronic relaxation or change in correlation energies causes
Koopmans' theorem to break down leading to limited quantitative success for
this approach.13 1)
The photoemission process comprises the following events:
1. Absorption of the photon by the bound electron, thereby increasing the
kinetic energy of the electron.
2. Ejection of the electron from the atom using some of its kinetic energy to
overcome the Coulomb attraction of the nucleus.
3. Simultaneous with 2, the electrons in the outer orbitals readjust (intra-
atomic relaxation) to a lower-energy final state, transferring this energy
to the outgoing electron.
Relaxation does not necessarily result in the formation of the ground state
of the ion. Simultaneous with the photoionization of an electron, there can be
excitation of one of the remaining electrons to an initially unoccupied orbital.
This phenomenon is called a "shakeup" process. The higher energy, low intensity
structure on XPS core-electron peaks can then be used to study various electronic
transitions occurring simultaneously with the photoionization. These satellite
features are found in the range from zero to 50 eV toward higher binding energy
than the main peak. Obviously, an electronic absorption at 50 eV ( = 404,000
cm-1 = 25 nm) is a very high energy absorption in the vacuum ultraviolet.
Photoemission from states belonging to a shell with an empty or partially filled
subshell will excite all states accessible by redistributing the electrons within that
shell that have the same symmetry as the state produced by the one electron
transition.
An example of such satellite XPS features appears in Fig. 16-8. Broad
features are seen toward higher binding energy than the two 0, peaks for
02. The features marked A, B, and C are "shakeup" peaks appearing as satellites
on the valence-electron peaks.
In ions with a partially filled outer shell, the coupling of the outer shell to
the core hole (all spin and angular momenta) will result in a set of final states
with distinct energies. The spectrum will exhibit fine structure (vide infra).
The "mixed-valent" species that was described in Chapter 10 [(NH 3),Ru-
(pyrazine)Ru(NH 3 ),]", has also been studied by XPS. 33 ) The XPS spectrum has
been reported to contain ionization peaks from two non-equivalent transition
metal ions. The carbon is peaks occur in the same spectral region as the metal
ionization peaks, and the conclusion rests on the ability to subtract these peaks
from the spectrum. The XPS technique has a very short time scale, on the order
of 10 -" sec. However, when a core electron is removed from a mixed-valent
compound, only discrete valences are obtained and fluctuations cannot persist
in the final state. Depending on the spin of the core electron removed, two
different spin states can result from the final ion. Thus, two "discrete peaks" in
C/200s C/200s
OXYGEN
0 02 0
Ols
2p 2p
2s -& 2 Gu2s
5000- ()() -500
W
Og du is
4000 -400 I
0
zz U
n 30
03000 300
Og 2s
(g 2p
2000- -200
C B A nu2p
ri)
ng2p
1000- -100
0 _0
eV 570 560 550 540 530 "60 50 40 30 20 10

BINDING ENERGY
FIGURE 16-8 XPS spectrum of core and valence electrons in 02 excited by Mg Ka

X-radiation. Two peaks are obtained for each fully occupied orbital, including the
oxygen is core orbital. 27 ) The peaks A, B, and C, and the peaks toward higher
binding energy from the 01, peak, are satellite peaks (vide infra). (From K.
Siegbahn, et al., "ESCA Applied to Free Molecules," North Holland/American
Elsevier, New York, 1969.)
the XPS do not distinguish mixed valence from delocalized systems. The reader
is referred to reference 34 for a critical discussion of the application of physical
methods to mixed-valent systems.
UPS
In the UPS experiment, vacuum ultraviolet radiation is used as the ionizing
radiation; usually this is provided by a helium [singly ionized, indicated as He(I)]
resonance lamp with an energy of 21.21 eV. Other discharge lamps have been
used. The energy of these lamps limits UPS to studies of valence electrons and,
in general, measurements have been mostly confined to gaseous samples. There
have been some reports of work on solutions 3" and solids.13 61
The UPS spectrum of a gaseous sample of N 2 is shown in Fig. 16-9. With
the He(I) source determining an ionizing limit of 21.21 eV, three vibrationally
structured photoionizations are seen (- 15.6, ~ 17.0, and ~18.8 eV). These can
be assigned to ionizations from the three highest-energy filled molecular orbitals
for N 2 , the 2a., r., and 3a, orbitals. The peaks have been assigned on the basis
of the observed vibrational structure. As an aside, it should be noted that an
XPS spectrum has the same three peaks (vibrational structure not seen because
of lower resolution) in addition to a peak at 37.3 eV for ionization from the 2a9
2 8
level as well as a single peak at 409.9 eV for both the 165 and 1, levels.( a)
In Chapter 3 the symmetry and construction of the molecular orbitals of
NH 3 were worked out. There it was found that the seven atomic orbitals in C ,
3
symmetry form a representation that is reduced to give three a, and two e
irreducible species. The eight valence electrons of NH 3 fill two of the a, and one
of the e molecular orbitals to give a ground state configuration of
... (2a 1 )2 (1e)'(3a1 )2
The only other filled orbital, the 1a, orbital, is essentially the nitrogen Is atomic
orbital. The He(I) (21.21 eV) spectrum is shown in Fig. 16-10, where vertical
ionizations are seen at 10.88 eV for the 3a, level and at 16.0 eV (first maximum)
for the le level. These assignments were made with the use of results from various
m.o. calculations. The more energetic He(II) (42.42 eV) source has been used to
observe the 2a, vertical photoionization at 27.0 eV. 3 s) In passing, it is of interest
to note the doubled or split character of the le peak at - 16 eV. This splitting
has been assigned to Jahn-Teller splitting in the ion resulting from a
(2a 1 )2 (le)3 (3a1 )2 configuration. The splitting is 0.78 eV. Other examples of Jahn-
Teller splitting have been reported.(2 8Sb
The vibrational structure on a UPS band tells something about the bonding
characteristics of the electron that is ionized. The ionization of a non-bonding
electron will result in an ion with the same internuclear distance as the parent
molecule. In this case the lowest energy (v = 0, v' = 0) vibrational peak will
dominate the spectrum (see Fig. 16-11). Referring to the N2 spectrum in Fig.
16-9, we see that ionization of electrons from either the 3a, level at 15.6 eV or
the 2a, level at 18.8 eV in each case gives essentially only one strong peak. These
levels are weakly antibonding and weakly bonding, respectively.
Photoionization of either a strongly bonding electron or a strongly anti-
bonding electron will result in changes in equilibrium bond distances, as illus-
trated in Fig. 16-11. In these cases, the most intense peak (i.e., vertical ionization)
3,000 r-
18.75(9)
A
I I I I I I
19.5 19.0 18.5 18.0 17.5 17.0 16.5 16.0 15.5 15.0
2P -4- -4- -4- -+-- 2P
0,15s
ISA ----- ------- Is
(B)
FIGURE 16-9 PES spectrum for a gaseous nitrogen sample. [From D. W. Turner
and D. P. May, J. Chem. Phys., 45, 471 (1966).]
will be at higher energy than the peak corresponding to adiabatic ionization.

This is exemplified by the n, peak for N 2 . In fact, the relative intensities (called
Franck-Condon factors) of the respective vibrational peaks can be theoretically
calculated with reasonable success.(3 7 ) The vibrational frequencies observed for
the ionized molecular state are also of further assistance in characterizing the
molecular orbital from which the electron is ionized. The ground state stretching
100-
FIGURE 16-10 PES

spectrum of a gaseous
sample of ammonia. (From
D. W. Turner, et al.,
"Molecular photoelectron
Spectroscopy,"
Wiley-Interscience, New
York, 1970.)
19 18 17 16 15 14 13 12 11 10
eV
FIGURE 16-11 Vibrational

states for the parent
molecule M and the ionized
molecule M+. (A) In this case
Ionized the electron ionized comes
Molecule
3 from a non-bonding orbital
2 and the two electronic
potential wells lie one above
V = 1
the other. The most intense
transition (indicated with an
arrow) is the V = 0<- V=0
transition. (B) The wells are
displaced (molecule
dimensions change) when
M the electron ionized comes
from either a bonding or an
antibonding orbital. In this
case the V'= 2<- V= 0
V 0 transition is the most intense
(A) (B) band seen.
frequency for N 2 is 2345 cm-'. Table 16-2 lists the stretching frequencies seen
for various N2 + ionized states. A comparison of each of these values with
the 2345 cm-' N 2 frequency leads to the implied m.o. characters of the lost
electron given in Table 16-2.
In addition to vibrational structure, other fine structure is seen in UPS
spectra. The first (i.e., lowest binding energy) peak in the spectra of CH 3 Cl,
CH 3Br, and CH 3I is, obviously, from the highest occupied molecular orbital,
and this is largely localized on the halogen atom (with some X-H anti-bonding
TABLE 16-2. Vibrational Structure in the

N2 UPS Spectrum
Implied
Peak v(N -N) M.O. Character
a,(2p) 2150 cm-' weakly bonding

n,(2p) 1810 cm-1 strongly bonding
a(2s) 2390 cm -1 weakly antibonding
and C-H bonding character.<3 s) Spin-orbit splitting is seen in this peak, where
the separation varies in the series: CH 3 I+, 0.62 eV; CH 3Br', 0.31 eV; and
CH 3 Cl', ~0 eV.(2"') These spin-orbit interactions agree with Mulliken's predica-
tions of 0.625, 0.32, and 0.08 eV, respectively.( 39 ) Many other such observations
of spin-orbit interactions have been noted. In fact, recent work at the high
resolution of 10 meV has detected spin-orbit splittings in NO of 0.012 to
0.016 eV.(40 )
A considerable number of small gaseous molecules have now been studied
with UPS and, for that matter, XPS. Some radical and excited state species have
also been studied. The three expected states (2p, 4 S, and 2 D) of 0 have been
detected. (4 An electrodeless microwave discharge was used to produce the
excited state species 0 2(0A.), and the photoelectron spectrum of this species
shows a vibrationally structured peak owing to formation of 02 1( 2fl 5 ). Compar-
ison with the peak corresponding to ionization of ground state 02(3X1 ) to the
same ionic state gives a value of 11.09 ± 0.005 eV for the adiabatic ionization
potential of O 2(A).-<4 2 ) Several other such species (e.g., SO and NF 2) have been
studied.
In comparison with diamagnetic compounds, paramagnetic species show
additional complexity in the UPS (and XPS) spectra. The oxygen molecule has
two unpaired ng electrons. The UPS spectrum of 02 is given in Fig. 16-12. The
photoionization of an electron from the partly filled, antibonding ng( 2p)
molecular orbital appears as the first peak in the UPS spectrum. Only one ionic
state is realized. On the other hand, photoionization of an electron from one of
the other filled molecular orbitals leads, in each case, to two electronic states of
the ion 02*. Thus, if an electron is ionized from the filled bonding n, level, the
unpairedelectron remaining in the nu orbital can be aligned either with or against
the two unpaired electrons in the antibonding nrg level. When the electron is
aligned with the two 7, electrons, there are three unpaired electrons, the total
spin S = 3/2, and the electronic state of the 02 + molecule is 4 fl. The other
alignment gives a 2 H electronic state for the 02 molecule. The 4Hu and 2 fIu
states of 02 + are at different energies, and thus there is a splitting of the nZ
orbital ionization peak. In Table 16-3, the observed features for the 02 molecule
from both UPS and XPS spectra are listed.
The XPS spectrum of gaseous 02 is shown in Fig. 16-8. The oxygen 1s peak
is also split, in this case by 1.1 eV. This splitting is not the difference between the
02 molecular orbitals 1so- and 1sag, which is calculated to be small. Recall that
no such splitting is seen in the case of N 2. Even further, the intensity ratio of
the two O, features is 2:1 and not the 1:1 ratio expected if the 1saulsa,
explanation applied. Again, as occurred in the valence orbitals, there is an
interaction between the unpaired electron in the 0,s level and the two unpaired
2 3
O?4E,
0$,4E,~ 0I, 0 - -0,X Ergv0
2''I
V' I.P. FCF V' I.P.
0 16 12
0 2029 0.18 0 18.17 1 1626 0 12.08
1 20.42 0.24 1 18 33 2 16.37 1 12.32
2 2055 022 2 18.45 3 16.49 2 12.54
3 20.67 0.16 3 18.58 4 16.60 3 12.73
4 20.78 0,12 4 18.71 5 16.72
5 20.88 0.04 6 16.83
6 20.96 0.03 7 16.93
8 17 04
FIGURE 16-12 Oxygen PES
spectrum. (From H. A. 0. Hill
and P. Day, "Physical
Methods in Advanced
Inorganic Chemistry,"
Interscience, New York, p.
88, 1968.)
10001
I I I I I I k
1 2 3 4 5 6 7 8 9 10
Electron energy (ev)
TABLE 16-3. Vertical 02 Ionization Data'
One-Electron Ionic Electronic

Molecular Orbital State XPS, eV PES, eV
2
ng p lg 13.1 12.10
4 17.0 16.26
U,2p
{4 18.8 18.18
2E, 21.1 20.31
2s
"u
{4E, 25.3 24.5 [viaHe(1I)]
2E 27.9
41 39.6
,2s 2Eg 41.6
4y 543.1
Oxygen Is 2y 544.2
"See reference 27.
electrons in the02 ' 7Eg antibonding molecular orbital. Ionization of even these
oxygen core electrons leads to two 02+ states, 2Y and 4Y, that differ appreciably
in energy.
The observation of such splittings is very interesting. One-center exchange
integrals dominate the energy difference between the quartet and doublet states.
Thus, interelectron repulsions (e2 /rij, where rij is the distance between the ith
and jth electrons and e is the charge on the electron) of the exchange type are
appreciable even between valence unpaired electrons and the oxygen core
676 chapter 16 Ionization Methods
electron. Such interelectron spin polarizations are found to be of importance in

explaining the shifts in nmr peaks for paramagnetic molecules (see Chapter 12).
Electron exchange splitting is also seen for core electrons of paramagnetic
transition metal complexes, as is described in a later section.
It is clear from the above discussion that the UPS spectra of relatively large
molecules are quite rich with information about ionization potentials, vibrational
quanta for the ionized molecule, spin-orbit interactions, Jahn-Teller splittings,
and electron exchange interactions. Unfortunately, there is frequently an over-
lapping of features, and broad peaks appear with no resolved vibrational
structure. As an example of a small molecule with many photoionization peaks,
Fig. 16-13 illustrates some of the spectral detail obtained14 11 for gaseous NO at
a resolution of 10 meV for a He(I) source and at 25 meV for a He(II) source. For
FIGURE 16-13 (A)

Photoelectron spectrum of
NO using the He 304-A line; (A) ' ' I"" " *""Ut* i"
three small peaks marked 24 22 20 18 16
'a" are due to the He 320 A eV
line; recording time was 75
h. (B) Photoelectron
spectrum of the B' 'I+ state
of NO using the He 304 A
line; the peak at "a" is
again due to the 320 A line.
(C) Photoelectron spectrum (B) I I
of NO using the 584 A line. 24.0 23.5 23.0 22.5
Resolution 10 meV.
Deconvolution of the peaks is
indicated by the dotted lines.
[From 0. Edqvist, L. Asbrink,
and E. Lindholm, Z.
Naturforsch., 26a, 1407
(1971).]
3
11- 1A 1- 1 1Z- 1A 3Z-
1 1 -1
18.8 18.7 18.6
eV
16-12 SIMS (Secondary Ion Mass Spectrometry) 677
the fine points of assignment the reader is referred to reference 40. Figure 16-13
shows that both exchange and spin-orbit splittings have been resolved.
16-12 SIMS (SECONDARY ION MASS SPECTROMETRY)
Secondary ion mass spectrometry(43' 4 4 ) employs an energetic (>4 keV) beam of

primary ions (e.g., 4 0Ar*, 2 0 2 Hg+, 160 ) to collide with surface species and
dislodge them as secondary ions. Upon collision with the surface, the primary
ion is implanted in the solid and its kinetic energy is transferred to atoms in the
condensed phase. This energy causes desorption (sputtering) of a surface species
producing gas phase ions. The mass spectra of these secondary ions are then
detected. Both positive ion and negative ion spectra are detected. Figure 16-14
shows both types of spectra that result from Ar + bombardment of an aluminium
surface.
10 -10
- Al' 101
Mg1
2
10 - A1 . A1*
12 - Na* K+ 2
1
: 10 Ar* A 2
H AIO + AIOH + - F] i AlO
l13 y 11 13 -0H c/
Al~Q~ AO2
l0-14 I.A 2 H Alt

11 1 4
C4
0 20 40 60 80 100 120 0 20 40 60 80
Mass number/charge (rn/e) -sC(A) Mass number/charge(r/e-
(A) (B)
FIGURE 16-14 SIMS spectrum. Secondary ion intensity versus mass number/ion
charge of an aluminum target under Ar bombardment. (A) Positive secondary ion
spectrum. (B) Negative secondary ion spectrum.
The escape depth for sputtered particles ranges from the surface to greater
than 20 A, depending on the target material and energy of the primary ion. The
predominant secondary ion species observed in SIMS are singly charged atomic
and molecular ions. As in mass spectroscopy, the isotopic composition aids in
identification of the fragments, which in turn leads to information about the
chemical composition of atomic layers at or near the surface. This is referred to
as static SIMS. By using ion etching (high primary ion current densities),
underlying layers can be exposed and the SIMS determined as a function of
depth. This is referred to as dynamic SIMS. In the dynamic SIMS mode, the
chemical composition of the material in the deeper layers is changed by the
prebombardment of the surface. In addition to chemical reactions, the more
volatile components can be "boiled" off. Static SIMS minimizes this problem
with negligible surface perturbation by the bombardment but provides informa-
tion only about processes occurring on the surface. Even in static SIMS
perturbation of the emission region before the secondary ion particle is emitted
cannot be excluded.("') SIMS instruments are available that provide lateral or
xy information about the surface. Ion microprobes provide this information
through probe imaging. Other instruments provide a direct image of the surface.
Since the detection limits are very low (1 ppb for many elements), surface
contamination by hydrocarbons, Cu, CO2 , H 2 , N,, 02, and H,0 is common.
The surfaces close to the sample can also be contaminated by sputtered material.
An additional complication is decomposition of the surface material by the high
energy primary ion source, thereby producing materials that are not actually on
the surface. The reader is referred to references 43 and 44 for a discussion of
experimental problems.
SIMS is used to identify surface species and also to study the dynamics of
surface processes (e.g., diffusion and corrosion). The catalytic decomposition and
synthesis of ammonia on iron metal has been studied.(4 ' The surface species
present on the commercial silver catalyst used to convert ethylene to ethylene
oxide have also been investigated.(4 6) The reader is referred to references 47, 48,
and 49 for reviews of SIMS applications.
In molecular SIMS, the primary beam expels molecular secondary ions from
samples of organic molecules supported on the surface. The molecular ions can
dissociate to give fragments that contribute to the SIMS spectrum. The region
just above the surface, called the selvedge, has a relatively high pressure where
ion-neutral molecule association reactions can occur. For example, the SIMS
spectrum of a silver film studied in the presence of benzene vapor produces
(AgCH,)* in SIMS. In the vacuum chamber, unimolecular dissociation reac-
tions occur in the course of traveling to the analyzer.
Molecular species are converted into secondary ions via three distinct
processes listed in order of efficiency: direct sputtering, cationization (or anioniza-
tion), and electron loss or gain. In sputtering, momentum transfer from the
primary ion dislodges organic cations and anions from the solid to the gas phase.
Organic dications in solids have been vaporized as gaseous doubly charged ions.
The limited degree of dissociation seen in SIMS spectra illustrates the low internal
energy of the ions formed.149 ) Cationization, in molecular SIMS, results from the
ion-molecule reaction during or following the formation of support metal ions
and the thermally induced evaporation of a supported organic molecule from
the surface. One of the complications in the interpretation of SIMS involves
distinguishing weak aggregates in a solid from aggregates formed in the selvedge
under ion bombardment. The ion-molecule complex formed in the selvedge can
fragment in the course of travelling to the analyzer much in the same way as the
unimolecular dissociations mentioned previously. The final process to be dis-
cussed involves electron transfer from or to the surface molecule generating M *
and M -. These ions can undergo unimolecular decomposition leading to
fragment ions. Examples of compounds that produce secondary ions by these
14
mechanisms have been reported. 11
Investigation of the chemistry that can occur when reactive primary ion
beams are employed represents a recent growth area for SIMS. Studies show
that N, + is a reactive primary beam.(5'0
16-13 LEED, AES, AND HREELS SPECTROSCOPY
LEED
In low energy electron diffraction (LEED), a monoenergetic beam of electrons
(10 to 500 eV) is incident on one face of a single crystal. About half of the electrons
are back scattered and the elastically scattered fraction hits a fluorescent screen.
A well-ordered surface results in a diffraction pattern. Surface crystallography is
16-13 LEED, AES, and HREELS Spectroscopy 679
carried out(") by finding the surface structure that optimizes the fit of an
estimated calculated pattern and the experimental one over a significantly large
range of diffraction conditions (electron energy, beam angles and number of
diffracted beams). In order to obtain information normal to the surface, the
variation in the intensity of a given spot in the diffraction pattern is studied as
a function of the incident beam energy. The theory involved in the calculation
of the experimental observables is complex.") Studies have enabled the compar-
ison of the surface structure with bulk structure and have shown the presence
of a terrace geometry involving stepped surfaces(5" (often one atom in height).
Oxidation or carbidization of the surface has been studied and the corresponding
change in surface structure determined. Reactivity of the stepped surfaces are
found to be different from those that are not stepped and different for various
types of steps.(5 2 )
A very interesting application of LEED involves determination of the
structure of ordered adsorbate-substrate structures. Adsorbate-substrate dis-
tances accurate to 0.1 A are claimed. A large number of surface structures for
gases (e.g., 02, H2S, C 2H4 , CH,, CO, CH 4 , C5HN) adsorbed on substrates
have been compiled.(5"
AES (AUGER ELECTRON SPECTROSCOPY)

In Auger spectroscopy (AES), the incident electron beam ionizes an atomic core
level electron. The energy released when a higher-energy electron falls back to
fill the core level is transferred to a valence electron, which is ejected into the
gas phase where it is detected as an AUGER electron with an electron spectrom-
eter.(53) The AUGER electrons appear as small peaks on a large background
of backscattered electrons including the elastically scattered primary electrons,
of LEED. The LEED optics can therefore be used to obtain the AUGER
spectrum and the two methods are often used in combination. The energy of the
AUGER electron varies with the surface element and its oxidation state. The
AUGER electron escape depth varies from ~4 to 25 A depending on the
material, so this is clearly only a surface method.(5 ') Diffusion, corrosion, surface
aggregation, and segregation can be monitored. The reaction of sulfur with 02 on
a Ni(1 11) surface was studied(5 4 ) by AES in conjunction with mass spectrometry.
The rate of growth of the oxygen peak and decrease in the sulfur peak was
monitored as a function of 02 pressure. The mechanism involved 02 adsorption,
reaction of adsorbed 02 with adsorbed sulfur to produce adsorbed SO2 , and
desorption of SO2 . The reaction to form adsorbed SO2 is rate controlling.
Adhesion, friction, and wear have been investigated showing changes in surface
2065
segregation of various components.(5 1 ) Changes in the surface composition of
heterogeneous catalysts have also been studied and correlated with catalyst
(5 1855
poisoning. 5,56)
C:
HREELS
- IiI IF ----
High resolution electron energy loss spectroscopy (HREELS) is another form of 0 1000 2000
electron beam spectroscopy. Vibrational excitation of surface atoms is induced (cmi)
by inelastic reflections of low energy electrons. The structure of the surface and FIGURE 16-15 HREELS
adsorbed species is determined. The HREELS spectrum of CO coadsorbed spectra of CO coadsorbed
with K on the Ru( 111) surface is shown in Fig. 16-15. The stretching frequencies with K on a Rh(111) surface.
F
for two different types of bound CO are seen at 1855 and 2065 cm-'. HREELS,
in combination with the foregoing techniques, has provided considerable insight
into heterogeneous catalysis. Studies of the hydrogenation of ethylene on
platinum, the hydrodesulfurization of thiophene on sulfided molybdenum, the
reforming of hexane, and the hydrogenation of carbon monoxide have been
summarized.(57 )
16-14 STM (SCANNING TUNNELING MICROSCOPY) AND

AFM (ATOMIC FORCE MICROSCOPY)
The resolution of microscopic techniques is determined by the size of the

measuring probe. With conventional microscopies, using light or electron beams,
the area resolved encompasses more than 10" atomic sites, producing an
averaged macroscopic picture of the surface. In 1982, the Nobel Prize in physics
was awarded to G. Binning and H. Rohrer""8 ) for developing the scanning
tunneling microscope-a method that provides the topography of surfaces with
atomic resolution.(58 )
In STM(5 9 ,60 >a sharp conducting metal tip is brought to within 5 to 10 A
of the sample. The tip traces the contours of the surface with atomic resolution
so it must be very sharp, ideally terminating in a single atom. At 5 to 10 A, the
wave functions of the sample and the tip overlap. If a bias is applied to the
sample, an electron tunneling current flows(61 > between the tip and the surface.
The electron flow can be in either direction depending on the sign of the bias
imposed. The tip can be moved in the x, y, and z directions using three
orthogonal x, y, and z piezoelectric translators. The tip moves in steps of
~1 nm (-three atoms) per volt applied to the translator. Moving the x and
y translators maps the surface while the z translator varies the tip-surface
distance. The tunneling current is very sensitive to the tip-surface distance,
typically changing by a factor of 2 for a 0.1 nm change. Differences of 1/100 of
an atomic diameter can be determined in the z direction. As one performs the
scan, the voltage into the z translator is changed to change the tip-surface
distance in order to maintain a constant current, typically 1 nA. Alternatively
the tip-surface distance can be kept fixed and the tunneling current measured.
With a tip the size of an atom, atomic resolution is achieved. Figure 16-16 is an
STM image showing carbon atoms in a sample of highly oriented pyrolytic
graphite. The dashed lines in the figure correspond to a scan in the x direction.
The voltage along z required to vary the distance and keep the current constant
is measured as a function of the voltage applied to move x. The image is made
by multiple scans changing the y direction for each.
The STM has been used in technological applications to guide the man-
ufacture of a spectral grating master(62 ) and in improving magnetic recording
FIGURE 16-16 An STM
heads.(6 3) By removing and depositing atoms, images with atomic dimensions
image of carbon atoms in a
have been made. By exciting individual atoms, nanoscale surface chemistry can
highly oriented pyrolytic
graphite sample. The image be induced.( 64 ) The surface chemistry of silicon has been explored and changes
is a series of line scans, that occur when chemical reactions take place monitored.(60 ) Reference 60 is
each displaced in the y highly recommended reading.
direction from the previous The application of STM is limited to conducting samples. Some success
one, displaying the path of has been achieved by studying conducting images of non-metallic substances,(65 )
the tip over the surface. by direct ejection of electrons into the conduction band,(66) and by using
nonlinear alternating current STM. 6 7 ) The atomic force microscope, AFM, can
image non-conducting surfaces.(68 ) A sharp tip, e.g., a small fractured diamond
fragment, attached to a spring in the form of a cantilever, traces over the sample
and senses the repulsive forces between the tip and the sample. Repulsive forces
usually in the range 10~6 to 10-9 N are recorded by measuring minute deflec-
tions of the cantilever. The spring deflection is measured by electron tunneling
with an interferometer by deflection of a laser beam off a mirror on the spring
or with a fiber-optic interferometer. Boron nitride has been imaged in this
way.( 6 9 )
16-15 INTRODUCTION EXAFS and XANES

Every element shows discontinuous jumps or "edges" in its X-ray absorption
coefficient corresponding to the excitation of a core electron (Fig. 16-17(A)). The 2p1s
energy of the edge depends on the element and the core electron excited. The K liii Ii i
edge (is state) is observed for elements between sulfur and cadmium and the L
edge (2s and 2p states) for elements of higher atomic numbers. Extended X-ray
absorption fine structure (EXAFS) employs high intensity X-ray beams (energies
to 40 keV; wavelength to 0.25 A) from a synchnotron source. The brightness is Absorption L2
edge L3 L,
as much as 10 orders of magnitude larger than that from conventional X-ray
tubes. A typical plot of the absorption coefficient as a function of the synchnotron
photon energy, E, is shown in Fig. 16-17(B). X-ray absorption spectroscopy 7 0 )
refers to the study of the structured absorption which is superimposed on the
edge.
The EXAFS region is ~50 to 1000 eV above the edge, whereas the X-ray
absorption near-edge structure (XANES) region refers to absorption within 50 eV
of both sides of the edge. The same principles govern absorption in both EXAFS
and XANES, but different information can be obtained.
The incident photon energy in EXAFS is much larger than the core-binding XANES EXAFS
Energy
energy leading to an excited photoelectron with significant kinetic energy. The
emitted photoelectron can be pictured as a wave propagating from the excited (A)
atom. Upon reaching a neighboring ligand, the wave front is scattered back FIGURE 16-17 (A)
toward the absorber. As the energy of the X-ray is increased, constructive and Schematic representation of
destructive interference of the outgoing and backscattered wave occurs near the absorption edges in X-ray
origin giving rise to the sinusoidal variation of the amplitude versus E, (Fig. absorption spectroscopy. (B)
16-17(B)), i.e., EXAFS fine structure. The amplitude and frequency of this Fine structure on a
absorption edge.
sinusoidal modulation depends on the type and bonding of the neighboring
atoms and their distances away from the absorbing atom. The scattering is similar
for atoms in the same row of the periodic table. Thus, C, N, and 0 are not
distinguished but N and S are. After considerable mathematical manipulation,
EXAFS provides the same compositional information as can be obtained from
X-ray crystallography with the advantage of being applicable to non-crystalline
systems. However, the mean free path damping factor limits information to -4 A
from the absorbing atom and does not provide information regarding the angular
arrangement of the surrounding atoms. It should be emphasized that EXAFS
measures a bulk property and is not a surface technique. If the sample is not
pure, a meaningless average EXAFS spectrum will result.
16-16 APPLICATIONS
Many of the above limitations turn out to be an advantage in the elucidation

of the coordination environment of metals and metal clusters in biological
systems. The molybdenum site in a nitrogenase enzyme was shown(71 ) to be
predominantly coordinated by sulfur ligands with another metal (inferred to be
iron) -2.7 A from a molybdenum center. Several model compounds were
synthesized with structures consistent with the EXAFS results. These models
contain oxygen and nitrogen ligands that could not be seen in the EXAFS of
the models in the presence of the sulfur ligands and thus are not expected to be
seen in the EXAFS of the enzyme. EXAFS was also used( 7 1) to show sulfur
coordination to the iron heme of cytochrome P-450, which is retained throughout
the catalytic oxidation cycle. EXAFS provides very accurate bond distances. It
improved(71) considerably on the iron-sulfur distances from an x-ray crystal
structure of the iron-sulfur protein, rubredoxin. The initial x-ray study reported
non-equivalent metal-sulfur distances. EXAFS showed them to be the same to
within 0.04 A. Other applications illustrating the advantages and limitations of
1
the technique have been summarized.(? )
REFERENCES 1. R. A. W. Johnstone, "Mass Spectrometry for Organic Chemists," Cambridge

CITED University Press, London, 1972.
2. M. R. Litzow and T. R. Spalding, "Mass Spectrometry of Inorganic and Organo-
metallic Compounds," Elsevier, New York, 1973, J. Lewis and B. F. G. Johnson,
Ace. Chem. Res., 1, 245 (1968); M. 1. Bruce, Adv. Organomet. Chem., 6, 273 (1968).
3. M. E. Rose and R. A. W. Johnstone, "Mass Spectrometry for Chemists and Bio-
chemists," Cambridge University Press, Cambridge, MA; 1982.
4. 1. Howe, D. H. Williams, and R. D. Bowen, "Mass Spectrometry: Principles and
Applications," 2nd ed., McGraw-Hill, New York, 198 1.
5. F. W. McLafferty, "Interpretation of Mass Spectra," 3rd ed., University Science
Books, Mill Valley, CA, 1980.
6. For brief reviews of recent developments in the field, see the annual summaries in
Analytical Chemistry.
7. a. G. M. Pesyna, R. Venkataraghavan, H. E. Dayringer, and F. W. McLafferty, Anal.
Chem., 48, 1362, (1976).
b. D. Henneberg, Adv. Mass Spectrom., 8, 1511 (1979); S. R. Heller and G. W. A.
Milne, Amer. Lab., 12(3), 33 (1980).
8. a. "Registry of Mass Spectral Data," E. Stenhagen, S. Abrahamsson, and F. W.
MeLafferty, Wiley, New York, 1974, 4 vols.
b. "Eight Peak Index of Mass Spectra," 2nd ed., Mass Spectroscopy Data Centre,
Aldermaston, Reading, U.K., 1974.
9. D. Williams, in "Determination of Organic Structures by Physical Methods," Vol. 3,
eds. F. C. Nachod and J. J. Zuckerman, Academic Press, New York, 1971.
10. H. M. Rosenstock and M. Krauss, in "Mass Spectrometry of Organic Ions," ed. F.
W. McLafferty, Academic press, New York, 1963.
11. K. Biemann and G. Spiteller. J. Amer. Chem. Soc.. 84. 4578 (1962).
12. J. H. Jones, Quart. Rev. Chem. Soc., 22, 302 (1968).
13. E. Hugentobler and J. Loliger, J. Chem. Ed., 49, 610 (1972).
14. J. Lederberg, J. Chem. Ed., 49, 613 (1972).
15. D. Maurer et at., J. Chem. Ed., 51, 463 (1974).
16. H. M. Bell, J. Chem. Ed., 51, 548 (1974).
17. J. Collard, Ph.D. thesis, U. of Illinois, Urbana (1975).

18. F. E. Brinckman and F. G. A. Stone, J. Amer. Chem. Soc., 82, 6235 (1960).
19. a. H. D. Beckey, A. Heindrichs, and V. Winkler, int. J. Mass Spectrom. Ion Phys.,
3, 11 (1970).
b. R. B. van Breeman, L. R. B. Martin, and A. F. Schreiner, Anal. Chem., 60, 1314
(1988).
20. T. A. Milne and H. M. Klein, J. Chem. Phys., 33, 1628 (1960).
21. R. T. Grimley, R. P. Burns, and M. G. Inghram, J. Chem. Phys., 34, 664 (1961).
22. J. Berkowitz, M. G. Inghram, and W. A. Chupka, J. Chem. Phys., 26, 842 (1957).
23. a. T. E. Sharp and J. R. Eyler, Int. J. Mass Spectrom. Ion Phys., 9, 421 (1972).
b. M. B. Comisarow, Anal. Chim. Acta, 178, 1-15 (1985).
c. J. D. Baldeschwieler, Science, 159, 263 (1968).
24. a. T. A. Lehman and M. M. Bursey, "Ion Cyclotron Resonance Spectrometry," Wiley,
New York, 1976.
b. R. T. McIver, Jr. Amer. Lab, 12(11), 26 (1980).
25. a. P. Kebarle and S. Chowdhury, Chem. Rev., 87, 513 (1987).
b. S. Chowdhury and P. Kebarle, J. Amer. Chem. Soc., 108, 5453-5459 (1986).
26. a. R. G. Keesee and A. W. Castleman, Jr. J. Phys. Chem. Ref. Data, 15, 1011 (1986).
b. M. T. Bowers, ed. "Gas Phase Ion Chemistry," Academic Press, New York, 1979,
Vols. 1 and 2.
c. P. Sharpe and D. E. Richardson, Coord. Chem. Rev., 93, 59 (1989).
d. J. Allison, Prog. Inorg. Chem., 34, 628 (1986).
e. B. S. Freiser, Talanta, 32, 697 (1985).
27. K. W. S. Chan and K. D. Cook, J. Amer. Chem. Soc., 104, 5031 (1982) (see opening
remarks).
28. a. K. Siegbahn, C. Nordling, G. Johansson, J. Hedman, P. F. Heden, K. Hamrin, U.
Gelius, T. Bergmark, L. 0. Werme, R. Manne, and Y. Baer, "ESCA Applied to
Free Molecules," North-Holland/American Elsevier, New York, 1969.
b. D. W. Turner, C. Baker, A. D. Baker, and C. R. Brundle, "Molecular Photoelectron
Spectroscopy," Wiley-Interscience, New York 1970.
29. S. Evans, "Handbook of X-Ray and Photoelectron Spectroscopy," ed. D. Briggs,
Heyden, London, 1977.
30. J. W. Robinson, "CRC Handbook of Spectroscopy," Volume 1, CRC Press, Cleveland,
Ohio, 1974, pp. 517-752.
31. a. I. Adams, J. M. Thomas, G. M. Bancroft, K. D. Butler, and M. Barber, Chem.
Commun., 751 (1972); W. E. Swartz, P. H. Watts, E. R. Lippincott, J. C. Watts,
and J. E. Huheey, Inorg. Chem., 11, 2632 (1972); M. Barber, P. Swift, D.
Cunningham, and M. J. Frazer, Chem. Commun., 1338 (1970).
b. D. T. Clark, D. Briggs and D. B. Adams, J. Chem. Soc., Dalton, 169 (1973).
32. D. A. Shirley, "Photoemission in Solids," Volume 1, Springer-Verlag, Berlin, 1978.
33. P. H. Citrin, J. Amer. Chem. Soc., 95, 6472 (1973).
34. A. Ludi, J. Amer. Chem. Soc., 106, 121 (1984).
35. H. Aulich, B. Baron, and P. Delahay, J. Chem. Phys., 58, 603 (1973).
36. D. E. Eastman, "Photoemission Spectroscopy of Metals," in "Techniques in Metals
Research VI," ed. E. Passaglia, Interscience, New York, 1971; T. E. Fischer, Surface
Sci., 13, 30 (1969); C. R. Brundle, "Surface and Defect Properties of Solids," Volume
1, eds. M. W. Roberts and J. M. Thomas, Specialist Periodical Reports, The Chemical
Society, London, 1972.
37. D. W. Turner and D. P. May, J. Chem. Phys., 45, 471 (1966).
38. H. Kato, K. Morokuma, T. Yonezawa, and K. Fukui, Bull. Chem. Soc. Japan, 38,
1749 (1965).
39. R. S. Mulliken, Phys. Rev., 47, 413 (1935).
40. 0. Edquist, L. Asbrink, and E. Lindholm, Z. Naturforsch, 26a, 1407 (1971).
41. N. Jonathan, A. Morris, M. Okuda, D. J. Smith, and K. J. Ross, "Electron Spec-
troscopy," ed. 6., A. Shirley, North-Holland, Amsterdam, 1972.
42. N. Jonathan, D. J. Smith, and K. J. Ross, J. Chem. Phys., 53, 3758 (1970).
43. J. A. McHugh, "Methods of Surface Analysis," Elsevier, Amsterdam, 1975, Chapter 6.
44. A. Berringhoven et al. "Secondary Ion Mass Spectrometers-Basic Concepts, Instru-
mental Aspects, Applications and Trends," Wiley, New York, 1987.
45. V. I. Shvachko and Ya. M. Fogel, Kinet. Katal., 7, 722 (1966).
46. A. Benninghoven, Surface Sci., 28, 541 (1971).
47. Ya. M. Fogel, Int. J. Mass Spectrom. Ion Phys., 9, 109 (1972).
48. A. Benninghoven, Surface Sci., 35, 427 (1973).
49. R. S. Day, S. E. Unger, and R. G. Cooks, Anal. Chem., 52, 557A (1980).
50. G. M. Lancaster, F. Honda, Y. Fukuda, and J. W. Rabalais, Chem. Phys. Lett., 59,
356 (1978).
51. G. A. Somorjai and L. C. Kesmodel, "Surface Chemistry and Colloids," Chapter 1.
52. B. Lang, R. W. Joyner, and G. A. Somorjai, Surface Science, 30, 454 (1972).
53. A. Joshi, L. E. Davis, and P. W. Palmberg, "Methods of Surface Analysis," Elsevier,
Amsterdam, 1975, Chapter 5.
54. P. H. Holloway, and J. B. Hudson, Surface Sci., 33, 56 (1972).
55. G. A. Somorjai, J. Catal., 23, 453 (1972).
56. M. M. Bhasin, J. Catal., 34, 356 (1974).
57. F. Zaera, A. J. Gellman, and G. A. Somorjai, Acc. Chem. Res., 19, 24 (1986).
58. G. Binning and H. Rohrer, Rev. Mod. Phys., 59, 615 (1987); contains the acceptance
speech.
59. P. K. Hansma, V. B. Elings, 0. Marti, and C. E. Bracker, Science, 24,2, 157 (1988).
60. P. Avouris, J. Phys. Chem., 94, 2246 (1990).
61. J. Bardeen, J. Phys. Rev. Lett., 6, 57 (1961).
62. Perkin Elmer Corp., Irvine, CA in reference 59.
63. Censtor Corp., San Jose, CA in reference 59.
64. E. E. Erichs, S. Yoon, and A. L. de Lozanne, Appl. Phys. Lett., 53, 2287 (1988).
65. M. Amrein et al., Science, 240, 514 (1988).
66. P. Avouris and R. Wolkow, Appl. Phys. Lett., 55, 1074 (1989).
67. G. P. Kochanski, Phys. Rev. Lett., 62, 2285 (1969).
68. G. Binning, C. F. Quate, and Ch. Gerber, Phys. Rev. Lett., 56, 930 (1986).
69. T. R. Albrecht and C. F. Quate, J. Appl. Phys., 62, 2599 (1987).
70. B. K. Teo "EXAFS: Basic Principles and Data Analysis," Springer-Verlag, New York,
1986.
71. a. S. P. Cramer, K. 0. Hodgson, W. 0. Gillum, and L. E. Mortenson, J. Amer. Chem.
Soc., 100, 3398-3407 (1978). S. P. Cramer, W. 0. Gillum, K. 0. Hodgson, L. E.
Mortenson, E. I. Stiefel, J. R. Chisnell, W. J. Brill, and V. K. Shaw, ibid., 3814-3819.
b. S. D. Conradson, B. K. Burgess, W. E. Newton, K. 0. Hcdgson, J. W. McDonald,
J. F. Rubinson, S. F. Gheller, L. E. Mortenson, M. W. W. Adams, P. K. Mascharak,
W. A. Armstrong, and R. H. Holm, J. Amer. Chem. Soc., 107, 7935-7940, (1985).
72. J. E. Hahn, K. 0. Hodgson, L. A. Andersson, and J. H. Dawson, J. Biol. Chem., 257,
10,934-10,941, (1982). S. P. Cramer, J. H. Dawson, K. 0. Hodgson, and L. P. Hager,
J. Amer. Chem. Soc., 100, 7282-7290, (1978).
73. R. G. Shulman, P. Eisenberger, W. E. Blumnber, and N. A. Stombaugh, Proc. Natl.
Acad. Sci. USA, 72, 4002-4007 (1975); D. E. Sayers, E. A. Stern and J. R. Herriott,
J. Chem. Phys., 64, 427-428, (1976); B. Bunker and E. A. Stern, Biophys. J., 19, 253-264
(1977). R. G. Shulman et al., J. Mol. Biol., 124, 305-321, (1978).
74. a. J. E. Penner-Hahn, "Metal Clusters in Biological Systems," ACS Symposium
Series, ed. L. Que, Washington, DC, 1988.
b. J. C. J. Bart, Adv. Catal., 34, 203 (1986).
Exercises 685
1. If the accelerating potential in a mass spectrometer is decreased in running a spectrum, EXERCISES

will the large or small m/e ratios be recorded first?
2. Sulfur-carbon 7r bonding is not as effective as nitrogen-carbon R bonding. In the mass

spectrum of HSCH 2CH 2 NH 2, would the m/e = 30 or m/e = 47 peak be more intense?
3. Refer to Fig. 16-4 and recall the discussion on the relation of charge delocalization
and stability of the positive ion. Explain why the mass 92 peak corresponding to
CH 2* is less intense for the isomer of ethyl pyridine with the

ethyl
N
group in the 3 position.
4. a. What m/e peak in the mass spectrum of CH 4 would you examine as a function of
accelerating potential in order to determine the ionization potential of the methyl
radical?
b. Write an equation for the appearance potential of the fragment in part (a) in terms
of ionization potential and dissociation energy.
c. In evaluating the dissociation energy of part (b) from thermochemical data, what
is wrong with using one-fourth value for the heat of formation (from gaseous carbon
and hydrogen) of CH 4 ?
5. UPS spectra for even relatively small molecules like butadiene are complicated (see
figures). Using Huckel calculations on butadiene, assign the peaks in the butadiene
spectrum (the peaks at 8.6 and 10.95 eV are impurity peaks.) Three vibrational
progressions (1520, 1180, and 500cm- 1) have been tentatively identified in the first
band at 9.08 eV. Discuss the vibrational structure on this band (both shape and
magnitude) with respect to your assignment. (Bands are seen in the infrared spectrum
of ground state butadiene at 1643, 1205, and 513 cm-'.)
11.34
21 20 19 18 17 16 15
300r
SI I
9.6 9.4 9.2 9.0

eV
6. The effects of coordination of small molecules to transition metals can be studied with
photoelectron spectroscopy. Spectra are given below for CO and W(CO)6 . Qualitat-
ively explain what observations bearing on the bonding in this complex are possible.
Take into account symmetry and energy considerations if applicable.
1000 r
14.01(8)
19.68(8)
16.53(6)
N2
i A~A
A JL-
I ff ii
20 19 18 17 16 15 14 13
Exercises 687
W(CO) 6
T
H20
C
0
21 20 19 18 17 16 15 14 13 12 11 10 9 8
eV
7. Given below are the UV photoelectron spectrum and m.o. diagram for ammonia.
Bonding (b), non-bonding (n), and antibonding (*) levels are indicated.
/I \
2p---\
-1S
2s-
10eb
oo
13 12 11
a. What phenomena give rise to the two bands centered at 11 and 16 eV.
b. What gives rise to the fine structure observed for each band? (Explain in one or
two sentences.)
c. Is the fine structure what you would predict, assuming the m.o. diagram given?
Why or why not? How might you justify any anomalies? Hint: consider the
shapes of the "before" and "after" species.
8. Describe which of the techniques described in this chapter you would use to:
a. determine the molecular weight of a non-volatile solid.
b. measure the C-O stretching frequency of CO adsorbed on a surface.
c. determine the oxidation state of a metal surface species after a metal film was
exposed to Cl 2 -
d. determine if the entire film in (C) had reacted with Cl2 -
e. determine the atoms in the coordination sphere of a metal ion in a metallo-protein.
9. Describe the differences and features in common for LEED, AUGER, and HREELs.
X-Ray Crystallography
JOSEPH W. ZILLER AND ARNOLD L. RHEINGOLD*
17
17-1 INTRODUCTION
X-ray crystallography is the most powerful and unambiguous method for the
structure elucidation of solids available to modern scientists. X-ray diffraction
has grown steadily since Max von Laue discovered in 1912 that a copper sulfate
crystal could act as a three-dimensional diffraction grating upon irridation with
X-rays. Early diffraction experiments were recorded on photographic plates or
film. The labor involved in determining a three-dimensional structure from these
early experiments could easily lead to one's thesis being based solely on one or
two structural determinations. The advent of modern high speed computers,
automated diffractometers, and powerful structure solution programs has allowed
X-ray diffraction to become widely accepted as a necessary standard technique.
Progress in X-ray diffraction is probably best traced by the number and
complexity of structures being reported each year. Information compiled in the
Cambridge Crystallographic Database") illustrates how rapidly structural in-
formation is growing (Table 17-1). It is clear that researchers are increasingly
relying on crystallography.
TABLE 17-1. Growth in the Cambridge Crystallographic Database

Year Structures Added Avg. Number of Atoms/Molecule Total Structures
1960 224 15.1 583

1970 1260 27.9 5577
1980 4324 38.6 30475
1988 7162 51.7 71619
1989 8005 53.1 82130
* This chapter was written by Joseph W. Ziller, University of California - Irvine and Arnold
L. Rheingold, University of Delaware. 689
690 Chapter 17 X-Ray Crystallography
The basic requirement to conduct an X-ray diffraction experiment is a single

crystal that can withstand exposure to X-rays of a given frequency for many
hours or days. Depending on the experimental conditions and the information
desired from the diffraction experiment (e.g., does one need to establish atom
connectivity or the absolute configuration of a resolved chiral molecule?), an
X-ray diffraction experiment may be carried out in as little as a few hours(2 ) or
take several days.
It is the intent of this chapter to familiarize the novice with general X-ray
diffraction principles, illustrate useful techniques for growing/handling/mounting
crystals, explain briefly the diffraction experiment, and illustrate how to interpret
and evaluate the results of a single-crystal X-ray diffraction study.
Principles 17-2 DIFFRACTION OF X-RAYS
In order to understand how the interaction of X-rays with single crystals yields
a diffraction pattern and ultimately (in most cases) a three-dimensional crystal
structure, it is necessary to know basic diffraction physics. A convenient and
conceptually easy way to describe diffraction of X-rays from crystals is to consider
that diffracted beams are "reflected" from planes in the crystal lattice analogous
to the reflection of an object from a mirror. The mathematical expression used
to describe how X-rays interact with crystals to produce a diffraction pattern is
given by Bragg's law:
nA = 2d sin 0 (17-1)
where n is an integer, 2 is the wavelength of the radiation, d is the perpendicular

spacing between adjacent planes in the crystal lattice, and 0 is the angle of
Incidence and "reflection" of the X-ray beam. Figure 17-1 shows how X-rays are
"reflected" from planes in the crystal lattice. Electromagnetic waves 1 and 2 strike
planes P1 and P2, respectively, at points A and B making the angle 0. For
diffraction to occur, it is necessary for the waves generated (1' and 2') to be
"in-phase" when "reflected " from P1 and P2. Constructive interference (in-phase)
of the waves emanating from points A and B occurs only when the path lengths
travelled are an integral multiple of the wavelength, n2. From Fig. 17-1 it is seen
that the path-length difference is 2d sin 0. When this equals nX, the Bragg equation
is satisfied and the resulting constructive interference produces a diffraction
maximum or "reflection."
Since the foregoing discussion considers diffraction of X-rays from lattice
planes in a unit cell, these planes must be designated in a consistent manner.
This is done by assigning Miller indices to the lattice planes. Miller indices are
represented by (hkl) values, which are universally used to represent indices. Miller
indices are also used to designate lattice points hkl that correspond to the (hkl)
family of planes. Each "reflection" of an X-ray from a crystal is assigned a unique
hkl value. (Note: when Miller indices are written in parentheses (hkl), they
designate lattice planes or crystal faces and when simply written as hkl, they refer
to a lattice point or "reflection"). Miller indices representing lattice planes and
lattice points are shown in Figs. 17-2(A) and 17-2(B), respectively.
17-3 Reflection and Reciprocal Space 691
'Diffraction vector'
direction
Angle of
ncidence
2 90 - 0* 90 - o 2' FIGURE 17-1 "Reflection"
of X-rays from crystal lattice
'j planes.
d \ 20
P2 d(Deviation from
P3 direct beam)
P3
Path differences
each d sin 0
17-3 REFLECTION AND RECIPROCAL SPACE
It is now necessary to relate diffraction from crystal planes to Bragg's law and
the direction of the diffracted beams. Rewriting Bragg's law in the form sin 0 =
nA/2(1/d), it is clear that sin 0 is inversely proportional to the d-spacing in the
crystal lattice. As a result, structures having large d-spacings exhibit compressed
diffraction patterns and structures with small d-spacings show expanded patterns.
It would be more convenient to express a direct relationship between sin 0 and
d. This is accomplished by construction of a reciprocal lattice in which 1/d is
directly proportional to sin 0. A reciprocal lattice can be constructed from a
direct lattice as shown in Figs. 17-2(A) and 17-2(B).
From the origin 0, lines P, Q, and R are drawn normal to the direct lattice
planes (hkl). Points along the normals are marked off at distances I/d,, from 0
where dhkl is the d-spacing between lattice planes. The resulting reciprocal lattice
for a primitive (P) monoclinic cell is shown in Fig. 17-2(b). The mathematical
relationships between direct and reciprocal lattices can be found in reference 3a.
The relationship of the direct and reciprocal lattices is shown below for a
monoclinic cell.
The reciprocal lattice parameters (a*, b*, c*, #*, and V*) are determined
from the direct lattice parameters according to Table 17-2. Figure 17-3 shows
0(001)2
1 L1 ,, xs0 " a 1011 00
101
P10 ' 2--*00 110
(100)102
103 axis
202
203
300 204
301
(A) 3 303
V 304
x*axis (B) 304
FIGURE 17-2 (A) Direct and (B) reciprocal lattices. (Reproduced with permission
from M. F. C. Ladd and R. A. Palmer, "Structure Determination by X-Ray
Crystallography." 2nd ed. Plenum Press, New York.)
TABLE 17-2. Relationships between Direct and Reciprocal Lattices for the
Monoclinic Cell
a* = 1/(a sin fl) a = 1/(a* sin f)2=y=*=y*=903
b* = 1b b = 1/b* #* 180 -#
c* = 1/(c sin #) c = 1/(c* sin b*) V*= 1V = a*b*c* sin#*
V = 1/V* = abc sin #l
the reciprocal lattice outlined with dashes and the direct cell with solid lines. The
reciprocal lattice may be used to illustrate how Bragg's law is satisfied.
Figure 17-4 shows the a*c* section of a reciprocal lattice of a crystal bathed
in X-rays with wavelength ). The X-ray beam (line XO) passes through the
reciprocal lattice origin 0. The circle with radius 1/ is centered at point C such
that the origin 0 is located on its circumference. Point P is a reciprocal lattice
point at any convenient point on the circle and angle OPB is a right angle because
it lies within the semicircle. The following equations are derived from Fig. 17-4(a):
OP OP
sin OBP = sin 0= - -- (17-2)
OB 2x
sin 0 = (OP/2)). (17-3)

y
Because P is a reciprocal lattice point, l/dk, is the length of OP. Substitution
FIGURE 17-3 Direct (solid) and rearrangement yields sin 0 = 1./ 2 dhk or Bragg's law with n = 1, i.e., 1; = 2d
and reciprocal (dashed) sin 0. Since Bragg's law is satisfied, reflection occurs. Figure 17-4(B) shows that
monoclinic cells. the reflecting plane is perpendicular to OP and therefore parallel to BP because
(Reproduced with permission OPB is a right angle. The angles 0 that the incident and diffracted beams make
from G. H. Stout and L. H. with the reflecting plane must be equal for constructive interference and thus the
Jensen, "X-Ray Structure diffracted beam OD is at the necessary 20 angle to the incident beam. When Fig.
Determination. A Practical
17-4(a) is constructed in three dimensions, a*c*, a*b*, and b*c*, the resulting
Guide." 2nd ed. John Wiley
& Sons, New York, 1989.)
sphere is called the "sphere of reflection." Whenever a point on this sphere satisfies
the foregoing arguments, the Bragg condition for reflection is satisfied and a
diffraction maximum can be measured.
Plane
FIGURE 17-4 (A) Reciprocal lattice; (B) direct plane. (Reproduced with permission
from G. H. Stout and L. H. Jensen, "X-Ray Structure Determination. A Practical
Guide." 2nd ed. John Wiley & Sons, New York, 1989.)
17-5 X-Ray Scattering by Atoms and Structures 693
17-4 THE DIFFRACTION PATTERN
The diffraction maximum or X-ray data collected from a single crystal yields an
intensity, Miller indices, and the diffraction angle 20 for each "reflection." The
intensity of the reflection depends on the nature and arrangement of atoms in
the unit cell and 20 is dependent only on the dimensions of the crystal lattice. 82
0( 0
As previously stated, for reflection to occur, Bragg's law must be satisfied, but
it must be recognized that atoms are not usually arranged in a unit cell such
that all atoms lie directly on lattice planes. The "reflections" observed are
composed of the sum of waves scattered from different atoms at different positions
in the unit cell. Actually, all atoms (in fact, all electrons) contribute to the intensities (A)
of all reflections.
Superposition of waves is the method of combining different waves to
generate the resultant wave observed as the "reflection" or hkl. This principle
states that the resultant wave amplitude in a given direction is obtained by
summing the individual waves scattered in that direction. A vector representation
of this superposition is shown in Fig. 17-5. The magnitude IF I and phase angle
a may be calculated for the resultant F.The formulas may be found in reference 3a.
FIGURE 17-5 (A) Vector
representation of the
superposition of waves f1,
f2, and f3. (B) Resultant
17-5 X-RAY SCATTERING BY ATOMS AND STRUCTURES wave if F with phase angle Y.
(Reproduced with permission
The scattering of X-rays by the electrons in the crystal produces the diffraction from G. H. Stout and L. H.
pattern. The scattering power of a given atom is a function of the atom type, i.e., Jensen, "X-Ray Structure
number of electrons and (sin 0)/ . The scattering amplitude of an atom is Determination. A Practical
designated as the "atomic scattering factor" or "atom form factor" (f,). At 20 = 0 Guide." 2nd ed. John Wiley
(sin 0/ = 0) the scattering factor is equal to the total number of electrons in the & Sons, New York, 1989.)
atom. As (sin 0)/ increases, the scattering of X-rays from different electrons in
the atom will become more out of phase. Figure 17-6 illustrates the decrease in
the scattering factor for several atoms at increasing (sin 0)/. values.
Since the frequency of X-radiation is high, the diffraction time scale is very
short and all motion is frozen. However, since the structure determined will
represent the frozen atomic positions averaged over 10" to 1020 unit cells, the
structure will show the effects of thermal activity, and any differences in atomic
arrangement within the unit cell will be revealed as disorder.
The scattering power of a real atom is also affected by thermal motion or
vibration of the atom. Thermal motion causes the electron cloud to be more
diffuse. This phenomenon is usually associated with such factors as temperature,
how tightly bound atoms are in a molecule, or how rigid the molecules are packed
together in the crystal lattice. The correct expression of the structure factor for
a spherical or isotropic atom is:
f = foe- B(sin2 0,
(17-4)
where B is a temperature parameter related to the mean-square amplitude of

vibration (p2 ) by B= 87E2 2
Since we are interested in the scattering from a group of atoms that constitutes
the structure being investigated, we must define another quantity, the "structure
FIGURE 17-6 Scattering 20

factor curves for atoms as a Ca
function of (sin 01)). 15
from J. P. Glusker and K. N. 1 c
Trueblood, "Crystal
Structure Analysis. A
Primer." Oxford University 5
Press, Oxford, U.K., 1985.)
C1
0 0.2 0.4 0.6 0.8 1.0 1.2 1.3
sin /X-
factor" IF I or [Fhk1 1.The structure factor is proportional to the observed intensity

taken according to equation (17-5). (See reference 4).
I = K IFI 2 (L,)(Abs) (17-5)
K is an initially unknown scale factor, L, is the Lorentz and polarization correction

factor and (Abs) is absorption. The observed intensities or "reflections" are
converted into |Fb, I (F observed). These quantities are used to calculate electron
density maps from which atomic positions and ultimately the three-dimensional
structure are determined.
The electron density at any point in the unit cell is given by the Fourier
series in equation (17-6).
p(x, y, z) = -
Vh k
YZI Fhkl e- 2ii(h+k,+1) (17-6)
Satisfying this equation results in a three-dimensional mapping of the electron

density of the unit cell. An atom would be located at each value of p(x, y, z) in
the map with atomic coordinates x, y, z. Unfortunately, the foregoing equation
cannot be solved straightforwardly because only the magnitudes and not the
phases of the Fhkl values are contained in the X-ray data. The uncertainty of
knowing the phases of the measured F-values is known as the phase problem in
crystallography. Fortunately for the modern crystallographer (and especially for
the amateur crystallographer), there exist excellent computer programs that have
been very successful in cracking the phase problem (see reference 3a, p. 245 for
a review of the phase problem and reference 4, p. 248 for a list of computer
programs). Once the phase problem is solved, subsequent least-squares analyses
and difference-Fourier syntheses are used to refine atomic positions and locate
any missing atoms. (As a testament to the importance of the work that resulted
17-6 Crystal Growth 695
in direct calculation of phases, in 1985 the Nobel Prize in chemistry was awarded
to Jerome Karle and Herbert Hauptman.)
17-6 CRYSTAL GROWTH Crystals

The growing of high quality single crystals is often referred to as an "art." There
are many different methodologies and schemes used to grow crystals. The answer
to the question, "What is the best method for growing good crystals of this
compound?" is "The best method is the one that works!" Some of the standard
crystallization techniques are discussed here.
Most crystallization attempts begin with a saturated solution. It is usually
best to begin with as much solute as possible since working with only a few
milligrams of compound often results in very small or poor quality crystals.
Slow evaporation of the solvent is probably the most often used and simplest
crystallization technique. During evaporation it is best to minimize agitation or
vibration of the sample because this may cause crystals to "crash out" too quickly
or cause developing crystals to go back into solution. A small vial or flask fitted
with a septum into which a small needle is inserted makes an excellent and cheap
crystallization device. If the solution requires heat to dissolve the solid, it is often
best to allow the solution to cool slowly rather that immersing it in an ice bath
or putting it in a freezer. This can be done by simply placing the crystallization
vessel on a bench top. For very slow cooling, the vessel may be placed in a warm
oil bath and both the bath and sample allowed to reach ambient temperature
on their own. Alternatively, the temperature may be electronically decreased over
a period of time.
Vapor diffusion is another method for obtaining high quality crystals. Figure
17-7 shows an "H-tube" apparatus. A solute/solvent solution (Sl) is placed into
side A and a second solvent (S2) is placed into side B. S2 should be a solvent in
which the solute is less soluble. Upon mixing of the solvents by diffusion, crystals
should form since they should be less soluble in the mixed solvent system than
in the solvent (Sl). An H-tube is quite versatile since it may be easily heated or
cooled or used under an inert atmosphere such as nitrogen or argon for
air-sensitive compounds. S, S2
Liquid diffusion produces crystals at the zone where two solvents mix. In
FIGURE 17-7 H-tube
another approach, a solute/solvent solution (S1) is placed into the vessel and a
apparatus for growing
solvent of different density is "layered" on top of the mixture; Fig. 17-8 as the crystals by vapor diffusion.
two solvents diffuse, crystals should appear at the interface and grow as mixing
proceeds.
A rather elaborate method for growing crystals of low melting compounds
5
directly on an X-ray diffractometer has been developed.! Focused heat radiation
and simultaneous cooling in a capillary affords crystals that may be examined
on the same machine on which they were grown.
A method for obtaining good quality crystals for low-yield compounds or
for those that crystallize poorly is to complex the target molecule with something
such as triphenylphosphine oxide (TTPO), which itself forms high quality
crystals.(6 ) Basically, the approach involves either coupling or cocrystallizing a
compound that forms high quality crystals with one that does not. TTPO is a FIGURE 17-8
good proton acceptor, and when it couples with the desired crystal, it often Liquid-diffusion vessel.
imparts its crystallizing characteristics to the product.
17-7 SELECTION OF CRYSTALS
A crystal suitable for a single-crystal X-ray diffraction study must possess internal
order and be of the proper size and shape. Internal order means that a crystal
should be "pure"~ on the atomic level. A "twinned" crystal is one that has more
than one orientation of a crystal in the lattice, but is often confused with a
",multiple" or "grown-together" crystal. Twinned crystals may sometimes be
detected from the diffraction pattern. Observations to note are multiple spots
(reflections) lying close together on a rotation photograph or "split" reflection
profiles. It is often difficult to determine the unit cell of a twinned crystal and if
it is decided that a crystal is twinned, the best and least frustrating way to deal
with it is to throw it away and try another crystal. However, in some cases it is
a property of all specimens.
Crystals should not contain cracks or have satellite crystals stuck to them
and should be as equidimensional as possible; however, this is not a strict
Exhaust requirement. Many X-ray structures have been correctly determined and properly
~ refined using plates, needles, or even "potato-shaped" crystals. If the nature of
B the crystal permits, it may be ground to a sphere using a device such as that
A deitdin Fig. 17-9. Air (or N2, etc.) causes the crystal to tumble around in the
abrasive cavity producing a spherical or nearly spherical shape. If grinding is not
desired or successful (more likely), crystals can usually be cut and sized with a
razor blade.
Compressed A polarizing microscope may be used to determine whether a crystal is single
air or made up of several fragments. Under polarized light, a good crystal should
FIGURE 17-9 Air-driven appear uniformly light and dark with every 90 rotation. Finally, the diffraction
crystal grinder. (Reproduced pattern of a good crystal should show sharp single spots with no "tails" or severe
with permission from G. H. powder rings.
Stout and L. H. Jensen, The next concern we have is the size of the crystal. Several factors need to
"X-Ray Structure be considered. First, the crystal must be small enough to remain within the
Determination. A Practical uniform intensity of the primary X-ray beam since the entire crystal should be
Guide." Macmillan, New irradiated by the same intensity radiation. This uniformity is usually about
York, 1968.) 0.50 mm x 0.50 mm. Crystals larger than 0.50 mm in any direction should be
avoided or cut smaller. The optimum size depending on the diffracting power of
the crystal and its absorption coefficient is about 0.10 to 0.40 mm on an edge.
As stated previously, absorption of X-rays by crystals must be considered
when deciding how large a crystal to use or which radiation to use. The resultant
intensity I of the X-ray beam after passing through an absorbing crystal of
thickness (t) is given by the equation I e , where I is the incident beam
intensity and () is the linear absorption coefficient of the compound. Unfortu-
nately, since the linear absorption coefficient is a function of wavelength of the
radiation used and atomic number, it cannot be calculated until the makeup of
the structure is known. This does not pose many difficulties since it is not usual
for one to begin an X-ray study assuming a "light-atom-structure" and later
determine that the crystal was a heavy-atom complex. The linear absorption
coefficient for the complex is calculated by summing the mass absorption
coefficients (nep) for the individual atoms present according to equation (177).
p = (density) Y (mass %)(/p) (17-7)
Mass absorption coefficients for elements for several 'different types of radiation
are listed in the International Tables, Vol. III, pp. 161-165.
17-8 Mounting Crystals 697
Since the X-ray beam is only uniform to just over 0.50 mm, this should be
used as an upper three-dimensional limit. For "light-atom," organic-type crystals
with no heavy atoms, such as Br, I, or higher (in atomic number), crystals
approaching the upper limit may be used. It is usually necessary to use large
crystals for "light-atom" structures due to their low scattering power. With
organometallic or inorganic compounds that contain higher percentages of
transition-metals, lanthanides, actinides, or other elements that absorb X-rays,
crystals should be as uniform as possible and not exceed about 0.20 to 0.40 mm
on an edge.
17-8 MOUNTING CRYSTALS
Depending on the nature of the crystal and the apparatus to be used, there are
many different methods of crystal mounting that may be utilized. Once good
quality crystals are obtained and the size/shape requirements are met, the crystal
must be mounted on the diffractometer. Unlike traditional film methods, modern
diffractometers do not require that a crystal be oriented in any special direction
relative to the X-ray beam. It is therefore not necessary to be concerned about
the orientation of the crystal on the mounting device (goniometer head) but it is
preferable to mount crystals so that their most nearly cylindrical axis is aligned
with the z-axis of the goniometer head. A typical xyz style head is shown in
Fig. 17-10. The goniometer head allows the crystal to be moved in three directions
in order to center it in the X-ray beam. Typically, a crystal is secured to a glass
fiber or wedged in a capillary tube that is inserted into a brass or aluminum pin,
which in turn is inserted into the goniometer head. The assembly is then screwed
onto the instrument.
The method used to mount a crystal depends on the nature of the crystal X-axis X-axis lock Y-axis
and on the experimental conditions. The most often-used method to mount adjustment adjustment
air/moisture-stable crystals is to glue them to a glass fiber. Five-minute epoxy is
FIGURE 17-10 Goniometer
an excellent adhesive in most cases. A very small bead of epoxy will usually head. (Reproduced with
suffice and, in fact, large globs of glue should be avoided since some of the crystal permission from
may be obstructed making optical alignment of the crystal difficult and beam Nicolet/Siemens, User's
intensity will be diminished by random scatter. For crystals that are air sensitive, Guide. Siemens Analytical
moisture sensitive, or lose solvent, a capillary tube may be used. Crystals should X-Ray Instruments, Inc.,
be placed into the capillary tube under an inert atmosphere either in a glove box Madison, Wis.)
or glove bag. The capillary is then flame sealed to prevent decomposition.
Capillaries are suited for either room or low temperature data collection; however,
they are somewhat awkward when using a low temperature device that utilizes
a cold air or nitrogen stream to cool the crystal. A routine method of "oil
mounting" crystals for low temperature studies has been developed.(8) The crystal
is immersed in a thick hydrocarbon oil and attached to a glass fiber, which is
mounted on the end of an elongated brass or copper pin. The oil acts as an
adhesive and also as an insulator, keeping the crystal from decomposing until it
is quick-frozen in the cold stream. This method is very useful for mounting crystals
that decompose when separated from their mother liquor or solvent. Since
mounting crystals in a glove bag or glove box can be an exercise in futility, this
method allows one to even shape (with a razor blade) the desired crystal under
a microscope on the benchtop. Although "oil mounting" may not be applicable
for every crystalline sample, experience has shown that if one works quickly, even
the most delicate and sensitive crystals stand a very good chance of surviving
this procedure. (Note: at the University of California-Irvine, nearly all structural
determinations, approximately 100/year, are done using this method).
Methodology 17-9 DIFFRACTION EQUIPMENT
The principal instrument used in modern X-ray laboratories is the four-circle

diffractometer. Although more traditional film methods are still employed, these
are becoming less standard practice. Modern diffractometers do not require
alignment of crystals in anything but a random fashion. The orientation of the
crystal relative to the diffractometer geometry is determined automatically by a
computer.
The main advantages of the automated diffractometer are that the accuracy
of results is far greater than that available from film, the instrument runs
automatically via computer control, and the rate at which reflections are measured
allows researchers to employ X-ray diffraction as a "routine," albeit, unparalleled,
analytical tool.
There are four basic components of a four-circle diffractometer:
1. The X-ray source, which includes a high voltage generator and a sealed
X-ray tube or a high intensity rotating anode
2. The goniometer or "crystal orientor"
3. The detector, which is usually a scintillation counter
4. The computer, which is used for diffractometer control and data
manipulation.
17-10 DIFFRACTOMETER DATA COLLECTION
The heart of the diffractometer is the goniometer. The most widely used geometry
is that of the four-circle or Eulerian cradle. The Eulerian cradle consists of four
circles that together act to orient the crystal such that all possible reflecting planes
may be examined automatically. These circles are designated 2-theta, omega, phi,
and chi. Only three of the circles operate independently since 2-theta and omega
are usually coupled. Figure 17-11 shows that 2-theta and omega are mounted
about the same axis.
A structural determination is usually begun by mounting a crystal on the
instrument and taking a photograph, known as a rotation photograph, using
Polaroid film. The crystal is rotated on the spindle axis phi for several complete
revolutions taking about 10 minutes; the diffraction pattern is recorded on the
film. Film measurements are input to the computer which positions the goniometer
to locate and center (determine the angular settings of 2-theta, omega, phi, and
chi) on 10 to 15 reflections. An alternate method is to allow the computer to
randomly search for reflections on which to center. This method works well for
strongly diffracting crystals but does not allow one to visually inspect the
diffraction pattern of the crystal. Visual inspection is advised since problem
crystals may be discovered and dealt with by mounting a new crystal.
17-10 Diffractometer Data Collection 699
Detector
FIGURE 17-11 The Eulerian cradle

showing the four circles.
X-ray tube
-' x'
20
Once several reflections have been centered, automatic computer routines

are used with careful attention by the crystallographer to determine an orientation
matrix that relates the crystal's geometry to that of the diffractometer and the
crystal's unit-cell parameters (a, b, c, o, #l, y, and cell volume). Photographs along
the cell axes should be taken to determine the presence or absence of symmetry.
In most laboratories, this is also done on the diffractometer rather than by
traditional film techniques.
Data collection should begin only after accurate cell parameters are obtained
and after as much information about the crystal as possible is known. One should
always try to determine the crystal's space group, possibly from a fast limited
data set, prior to final data collection. The number of unique data points to be
collected depends on the type of radiation used, the Laue symmetry of the crystal,
and the quality of the crystal.
The total number of accessible reflections for a crystal is known as the
"limiting sphere" and depends only on the wavelength and is given without
further explanation as N = 33.5 (vol. direct cell/P). The volume of a molecule
containing a typical mix of light and heavy atoms may be approximated by the
remarkably consistant observation that the average atom occupies 9-11 A3
when dominated by heavier atoms, often using 18 A3 for just the non-hydrogen
atoms will prove more reliable. For a cell with volume = 3000 A3, the maximum
theoretical number of reflections for Mo radiation (A = 0.71073) is approximately
280,000, whereas for Cu (A = 1.5418) the number is approximately 27,000.
Fortunately, it is not necessary to collect all of these reflections. According to
Friedel's law, hkl = hkl, therefore the maximum number of reflections necessary
will be one-half of the preceding numbers. For higher-symmetry crystal systems,
the number of unique data is reduced even further. For instance, if the crystal
system is orthorhombic, the symmetry is mmm (three mutually perpendicular
mirror planes) and only one octet of data needs to be collected since
hkl= hkl = hki = hki = hk-ik= hkl = hki

Typical hkl ranges for triclinic, monoclinic, and orthorhombic crystal systems are:
Oki -+ hkl, triclinic

001-+ hkl, monoclinic
000 -- hkl, orthorhombic
Another factor that determines how much data should be collected is the fact
that the intensity of reflections decreases with increasing (sin 0)/^. Data are
collected between a min and max 2-theta (or theta) limit. Typical values of
26(max) for Mo radiation are 45 to 550 (or 0.54 to 0.65 (sin O)/l) and the total
number of observable reflections is given by No N, x sin 3 0. (For the foregoing,
3
No = 280,000 x sin (25) = 21,135 for 20-max = 500.) Crystal quality, possibly
due to thermal motion, may dictate that there will be no usable data beyond the
500 limit.
A method used to improve data quality that is quickly becoming routine in
many laboratories is low temperature data collection. As already discussed, low
temperature data collection allows for easier handling and mounting of crystals.
Reflection profiles are sharpened and backgrounds are usually reduced. Improving
the signal-to-noise ratio aids in resolution of weaker reflections and those at
higher (sin 0)/ values. Thermal motion is also reduced, although not eliminated.
Decomposition can be reduced or may be completely eliminated for crystals that
are prone to decay in the X-ray beam. Scan speeds for data collection can be
increased resulting in less X-ray exposure and ultimately in a greater number of
data sets being collected.
17-11 COMPUTERS
Many crystallographic facilities are moving away from large mainframe com-
puters to mini, micro, or even personal computers. Most crystallographic software
can be run on a variety of computer platforms. The usual laboratory setup
includes one computer that is used for both data collection and data reduction.
Graphics terminals, color printers/plotters, and laser printers enable one to
produce publication quality output without the necessity of employing specialized
services such as a draftsperson.
Some Future 17-12 AREA DETECTORS

Developments The field of protein crystallography is growing steadily. Because of the necessity
to determine the structures of large macromolecules, area detectors were
developed. Standard diffractometer practice is to collect one reflection at a time
and record the intensity using some type of scintillation counter. Area detectors
are able to measure the intensities of many reflections simultaneously. This is of
particular importance since it is necessary to measure a far greater number of
reflections for macromolecules. Also, measuring many reflections at once results
in less decay of the crystal. This is obviously a major concern when exposing
17-15 Crystal Classes 701
biological samples to X-rays. Future advances in this area will result in the
elucidation of larger and more complex crystal structures. As the resolution of
area detectors improves, they may also provide advantages for small-molecule
work as well.
17-13 X-RAY VERSUS NEUTRON DIFFRACTION
X-ray and neutron diffraction can be used as complimentary techniques because

of both similarities and differences between them. Whereas X-ray diffraction relies
on scattering by electrons, neutrons are scattered by nuclei. One of the most
common uses of neutron diffraction is to locate hydrogen atoms with high
accuracy because the degree of scattering by nuclei does not vary significantly
with atomic number. In fact, because hydrogen has a negative scattering factor
for neutrons, whereas that for deuterium is positive, they are easily distinguished
from one another. A disadvantage of neutron diffraction is that much larger
crystals are required than for an X-ray analysis.
17-14 SYNCHROTRON RADIATION
A highly specialized type of X-ray source is available through the use of particle
accelerators. Synchrotron radiation is produced by high energy electrons in a
storage ring when a magnetic field is applied, causing their path to bend. The
energy emitted is 100 to 104 times as intense as a conventional X-ray tube. The
wavelength of synchrotron radiation can be tuned for a particular experiment,
typically in the 0.8 to 2.0-A range desired by crystallographers. The main
disadvantages of synchrotron radiation is that there are a limited number of
installations and expenses are high.
17-15 CRYSTAL CLASSES Symmetry and

A unit cell is designated by the six lattice parameters a, b, c, a, f, and ,. These Related Concerns
correspond to the axial lengths and the angles between them. a is the angle
between b and c, # is the angle between a and c, and 7'is the angle between a
and b. (See Fig. 17-12.) The unit-cell volume, V may be obtained from
a
0 b FIGURE 17-12 Monoclinic
and orthorhombic unit cells
----- showing the axial lengths
and angles between them.
c c
Monoclinic cell Orthorhombic cell
(A) (B)
V = abc(1 - cos 2
- cos 2 # cos 2 y + 2coscos flcos y)1 2
The values of the six parameters may be either unique or they may be constrained
due to the symmetry of one of the seven crystal classes listed in Table 17-3.
The symbol # should be read "not constrained to equal," since it is possible for
TABLE 17-3. The Seven Crystal Classes
No. Independent Lattice

Crystal System Parameters Parameters Symmetry
triclinic 6 a# b # c; 2 =A#: 1
monoclinic 4 a# b # c; a = 90 ; f# 90 2/m
orthorhombic 3 a# b = c; 2 = =+ 90 mmm
tetragonal 2 a= b # c; = # == 90 4/mmm
rhombohedral 2 a= b c; =#= y # 90 3m
hexagonal 2 a= b #c; = 90 ; y 120 6/mmm
cubic 1 a= b c; 2 # = 7 = 90 m3m
Cubic
P I F
Tetragonal -- c - a
a a i a
Hexagonal P Rhombohedral R
Orthorhombic
P C I F
Monoclinic
a a a
P C Triclinic P
FIGURE 17-13 The fourteen Bravais lattices distributed among the seven crystal
systems. The primitive lattice (P) has 1 equivalent point per unit cell at (000); the
body centered lattice (C) has 2 equivalent points per unit cell at (000; 110); and the
face-centered lattice has 4 equivalent points per unit cell at (000; O"; '01; "0). All
equivalent lattice points are related to each other by pure translation and have
identical environments.
17-16 Space Groups 703
accidental equalities to occur, but this is rare. Addition of face and body centering
to the seven crystal systems leads to the 14 Bravais lattices shown in Fig. 17-13.
17-16 SPACE GROUPS Unit

translation
The point-group operations of rotation, mirror reflection, and inversion discussed
in Chapter 1 define the 32 unique point groups. Combining the 32 point groups
with the 14 Bravais lattices result in 230 space groups that are a group of symmetry
operations that describe how molecules are arranged, pack, interact, etc., in the
unit cell of a crystal lattice. It is now necessary to define two additional
FIGURE 17-14 Three-fold
"space-symmetry operators" that account for translational motion. These addi- screw axis. (Reproduced with
tional operators are the screw axis and the glide plane. permission from G. H. Stout
An n-fold screw axis is produced by combining a rotation axis and a and L. H. Jensen, "X-Ray
translation parallel to this axis. The rotation is (360/n) and the translation is a Structure Determination. A
fraction (m/n) of a unit translation. The integer n may take the values 1, 2, 3, 4, Practical Guide.'' 2nd ed.
or 6. The screw axis defined by the symbol 3, means that a point is rotated John Wiley & Sons, New
around the axis by 120 (360/3) and then translated by one-third of a unit York, 1989.)
translation. This is shown graphically in Fig. 17-14. The points represented by
2 and 3 result from successive 31 operations. Point 1' is indistinguishable from
1 since it is the result of a 3/3 or unit translation.
A glide plane is the result of combining a translation parallel to a mirror
plane with subsequent reflection across the mirror.The translation is either along
a cell edge or a face diagonal. Glide planes are designated as a, b, c, n, or d
depending on the translation. An a-glide is shown in Fig. 17-15. The translation
for glides a, b, and c is one-half of the axial length, for n-glides it is one-half of
bA c
the diagonal length. Diamond or d-glides occur for centered lattices and
FIGURE 17-15 Glide plane
correspond to a one-fourth diagonal translation. Since all glide planes translate
a. (Reproduced with
either by one-half or one-fourth, after either two or four glide operations, the permission from G. H. Stout
resulting point corresponds to a unit translation. A glide operation inverts and L. H. Jensen, "X-Ray
chirality; a screw does not. Structure Determination. A
In addition to screw axes and glide planes, centering operations may also Practical Guide." 2nd ed.
be present. P or primitive lattices have no centered lattice points. A, B, and C John Wiley & Sons, New
lattices have lattice points centered on their respective cell faces. I lattices have York, 1989.)
a lattice point at the center of the cell and F lattices have lattice points on all P2, (2)
faces. Centered lattices are shown in Fig. 17-13. 0 4- b
All 230 space groups are compiled in the International Tables, Vol i(7b)
Equivalent positions, special positions, and reflection conditions for each are
listed. The reflection conditions observed for the diffraction data are used to
determine the space group for a given crystal structure. Unfortunately, several
space groups have identical reflection conditions (usually referred to as systematic
absences) and are not distinguishable from one another solely on this basis. Table P21 , equivalent positions
17-4 lists symmetry elements and their associated systematic absences. This table (1 ) x, y, z; (2) x, y + -1 z
may be used to determine the space group of a crystal structure based on the
fact that symmetry elements give rise to specific systematically absent reflections. FIGURE 17-16 Equivalent
position in space group P21.
The equivalent positions (fractional coordinates where two identical molecules
are located in the unit cell) for the monoclinic space group P21 are illustrated in from G. H. Stout and L. H.
Fig. 17-16. The 2, screw axis relates every point at xyz to an equivalent point Jensen, "X-Ray Structure
at -x, y + ', -- z. The screw axis is along b such that the equipoint -x, y + %, Determination. A Practical
-z results from rotation about b and then translation of one-half of a unit cell Guide." 2nd ed. John Wiley
(the sign of y remains the same, whereas x and z become negative). & Sons, New York, 1989.)
TABLE 17-4. Symmetry Elements and Reflection Conditions

Affected Condition for Systematic
Symmetry Element Reflection Absence of Reflection
2-fold screw (2) a hOO h = 2n + 1 = odd

4-fold screw (42) along b Ok k = 2n + 1
6-fold screw (63 )l c 001 1 = 2n + 1
3-fold screw (3s, 32) along c" 001 l = 3n + 1, 3n + 2,
6-fold screw (62, 64) i.e., not evenly divisible by 3
4-fold screw (4k, 43) along a hOO h = 4n + 1, 2, or 3
b Ok k= 4n+ 1,2, or 3
c 001 1= 4n+ 1,2, or 3
6-fold screw (61, 65) along c" 001 1 = 6n + 1, 2, 3, 4, or 5
Glide plane perpendicular to a
translation b/2 (b glide) Ok1 k= 2n + 1
c/2 (c glide) I= 2n + 1
b/2 + c/2 (n glide) k+ l = 2n + I
b/4 + c/4 (d glide) k+ l = 4n + 1, 2, or 3
Glide plane perpendicular to b
translation a/2 (a glide) hOl h= 2n + I
c/2 (c glide) I= 2n + 1
a/2 + c/2 (n glide) h+ I = 2n + I
a/4 + c/4 (d glide) h+ I = 4n + 1, 2, or 3
Glide plane perpendicular to c
translation a/2 (a glide) hk0 h = 2n + 1
b/2 (b glide) k = 2n + 1
a/2 + b/2 (n glide) h + k = 2n + I
a/4 + b/4 (d glide) h + k = 4n + 1, 2, or 3
A-centered lattice (A) hk1 k + I = 2n + 1
B-centered lattice (B) h + l = 2n + I
C-centered lattice (C) h + k = 2n + I
Face-centered lattice (F) h + k = 2n + 1 i.e., h, k, I not
h + I = 2n + 1 allevenorall
k+ l = 2n + I odd
Body-centered lattice (1) h + k - I = 2n + 1
"Note that in the crystal classes in which 3- and 6-fold screws occur as cell axes, these are
conventionally assigned to be c, so only the 001 reflections need be considered.
17-17 SPACE-GROUP DETERMINATION
Despite the possible need to choose from among 230 options, in fact, most
molecular species crystallize in just a few different space groups: PT, P21/c, C2/c,
P21 21 11, Pbca, and Pnma account for about 85% of all structures.
The following two examples will illustrate how to determine the space group
of a crystal from the diffraction symmetry and systematic absences. The diffraction
symmetry may be determined using photographs and the systematically absent
reflections may be determined from the intensity data. (Refer to Table 17-4 for
the symmetry element/systematic absences relationships.)
Case 1. The diffraction symmetry is 2/m (a twofold rotation axis with a
mirror plane perpendicular to it), indicating a monoclinic crystal system.
There are no reflection conditions consistent with a centering operation
so the lattice is primitive P. Reflections of the type h0l are absent when
I is odd and the OkO reflections are absent when k is odd. According to
17-18 Avoiding Crystallographic Mistakes 705
Table 17-4, the hOl absence is a result of a c-glide plane perpendicular

to the b axis and the OkO absence is indicative of a 21 screw axis along
b. The space group is uniquely defined as P21/c.
Case 2. The diffraction symmetry is 2/m indicating a monoclinic cell. The
only systematic absences observed are for hkl reflections when the value
of h + k is odd. This information is consistent with a C-centered lattice.
Since there is no screw axis or glide plane indicated, the only possible
space groups are C2, Cm, or C2/m. These space groups are indis-
tinguishable from one another and it may be necessary to examine each
by trial structure solution/refinement to determine which is the correct
space group.
Case 2 represents a situation where one must make a choice between
centrosymmetric and non-centrosymmetric space groups (C2/m is centric, Cm
and C2 are acentric). If a structure contains an equal mix of left- and right-handed
molecules, it must crystallize in a centrosymmetric space group. This is true even
if the molecule is asymmetric. The handedness or chirality of a crystal means
that a structure cannot be superimposed on its mirror image. If a crystal is known
to be chiral and resolved, it must not be centrosymmetric. When more than one
space group is possible, this information may assist one in determining whether
a centric or non-centric space group is most likely the correct choice.
17-18 AVOIDING CRYSTALLOGRAPHIC MISTAKES
The aforementioned problem of determining whether a crystal is best described

as centrosymmetric or non-centrosymmetric requires further discussion. Since
centric/acentric space group "pairs" (e.g., C2/m, Cm, C2) have the same systematic
absences, these are of no help in sorting out the ambiguity.
Consider the monoclinic space groups P21 /m and P2, that are centrosym-
metric and non-centrosymmetric, respectively. The only systematically absent
reflections are the OkO when k is odd (twofold screw axis perpendicular to b). If
a structure is solved in P21 /m, it may also be described in P21 with a doubling
of the Z value (The Z value refers to the number of molecules in the unit cell.
The Z value for P21 /m with one molecule in a general position is 4, whereas for
P2 1 it is 2.) If a structure has molecular symmetry such that a mirror plane relates
one-half of the molecule to the other, then it is possible (but not required) that
the molecule possesses the same crystallographic symmetry and be located on
the crystallographic mirror plane in P2 1/m. This structure could also be described
in P21, but all atoms would have to be located. (The discussion of molecular
versus crystallographic symmetry will illustrate the requirement to locate only
the symmetry unique atoms for naphthalene.)
Frequently, a structure will be described as ordered in a non-centrosymmetric
space group rather than as disordered in a centrosymmet'ric one.") When this
situation occurs, it may be more correct to describe the structure as centrosym-
metric and disordered, indicating that only the "average" structure has been
determined. Several structures have been "republished" after investigation of the
original work revealed that the space-group assignment was incorrect." )
An example of a structure that was originally described in a non-centrosym-
metric space group and later shown to be more correctly described as centric is
taken from reference 9. The structure of [(1,4,7,10-tetrathiacyclododecane) cop-

per(II)bis-perchlorate-monohydrate] (Fig. 17-17) was described in the non-centric
orthorhombic space group Pbc2. Potential problems that should have been
carefully considered were:
1. A pseudomirror plane perpendicular to the c-axis that would be
consistent with the mirror plane in Pbcm (centric).
2. Problems in least-squares refinement of heavy atom positions.
3. Refinement of S atoms with anisotropic temperature factors led to
unrealistic bond distances requiring isotropic refinement of these atoms.
(An anisotropic temperature factor is an exponential expression that
FIGURE 17-17 Isotropic accounts for the vibration of an atom. This vibration is described as an
thermal plot of ellipsoid rather than spherically symmetric or isotropic.) See Fig. 17-19
[(1,4,7,10- for an example of a molecule refined with anisotropic thermal parameters
tetrathiacyclododecane) and Fig. 17-17 for the isotropic plot of [(1,4,7,10-tetrathiacy-
copper(II) bis-perchlorate- clododecane)copper(II)bis-perchlorate-monohydrate].
monohydrate]. 4. A high R value of 12.70% (R values are quantities that are used to
compare observed and calculated structure factors. See Section 17-20.
5. Large ranges for the Cu-S and S-C distances 2.30(1) to 2.37(1) and 1.73(4)
to 1.9(4) A.
6. Short C-C distances 1.41(3) A.
Refinement in the centrosymmetric space group Pbcm yielded the following
significant results.
1. Statistically equivalent Cu-S [2.325(3)] and S-C [1.815(9)] distances with
esd's about one-half as large as the original.
2. A much lower R value of 6.8 % with all atoms anisotropic.
There is considerable disorder in the Pbcm structure; the three S-C-C-S
groupings are disordered across planar conformations and there is a high degree
of disorder for the two perchlorate ions (a common problem). it is clear that the
ordered Pbc2 structure is just one of many possible orientations of the structure
that is better described by the "average" Pbcm model.
The moral of the foregoing example is to examine each possible space group
carefully when a choice must be made between a centrosymmetric or non-
centrosymmetric model. Some of the following things to note are:
1. Does the molecule possess any molecular symmetry as a mirror plane,
inversion center, or rotation axis that may also be crystallographic?
2. How many molecules are in the asymmetric unit? What is the expected
and observed Z-value?
3. Are pairs of atom coordinates similar such that they may be related by
some symmetry operator?
4. Are interatomic distances and angles consistent within accepted ranges?
5. Are the shapes of thermal ellipsoids reasonable?
The best rule to follow when one has a choice of centrosymmetric or
non-centrosymmetric space groups is to test the centrosymmetric group first. If
the structure cannot be solved or refined properly then attempt the solution with
the non-centrosymmetric space group. If any of the foregoing problems are noted
or if the structure fails to refine well, then investigate the centrosymmetric group
again until it is clear that the structure can only be described in the non-
centrosymmetric space group.
17-19 MOLECULAR VERSUS CRYSTALLOGRAPHIC SYMMETRY
There are often relationships between a structure's molecular and crystallographic

symmetry. consider the case of naphthalene and pyrene (Figs. 17-18(A) and
17-18(B)). Both molecules are highly symmetric with planar structures. Viewing
Figs. 17-17(A) and 17-17(B), it is clear that both molecules possess mmm
molecular symmetry (three perpendicular mirror planes), inversion centers, and
rotation axes. Naphthalene may be described as having mirror planes (1), the
m C12 C11
0 C2 C4
C1 C3 C5'1 C1 4 C13 C10 C9
m -c C1 C15 C16 C8
C5 C3' C1' C2 C3 C6 C7
C4' C2'
m m C4 C5
mm
(A) (B)
FIGURE 17-18 Ball-and-stick drawings of (A) naphthalene and (B) pyrene.
plane of the molecule, (2) a mirror along C3-C3', which would relate each half
of the molecule to itself, and (3) a plane bisecting the Cl-C5, C3-C3', and Cl'-C5'
bonds. There are also twofold rotation axes perpendicular to these mirrors that
relate halves of the molecule. In addition, there is an inversion center midway
between C3-C3' that interconverts the two rings. Whereas naphthalene exhibits
a high degree of molecular symmetry, the only observed crystallographic
symmetry is the inversion center between C3-C3'. Table 17-5 lists atom
coordinates for the inversion-related atom pairs. The inversion center is located
at (0, 0, 0). When solving the structure of naphthalene, only one-half of the total
number of atoms needs to be located and refined since the other half is related
by symmetry. All of the aforementioned noted molecular symmetry elements are
also present for pyrene. This would lead one to expect that at least some of these
elements are present in the crystal structure; however, pyrene has no crystallo-
graphic symmetry and all atoms must be located and included in the refinement.
These two examples should illustrate the fact that molecular symmetry does not
necessarily lead to crystallographic symmetry.
17-20 QUALITY ASSESSMENT
X-ray diffraction experiments produce results that are usually believed implicitly.
After all, there is a picture of the molecule so it must be correct and well done.
Although most X-ray studies are accurate, one should remember that the derived
TABLE 17-5. Atomic Coordinates and Their Inversion-

related Pairs for Naphthalene
Axes
Atom x y z
C1 0.08232 0.01856 0.32836

C2 0.11295 0.16382 0.22289
C3 0.04799 0.10518 0.03714
C4 0.07656 0.25183 -0.07582
C5 -0.01320 -0.19021 0.25460
HI 0.12420 0.05890 0.45540
H2 0.17870 0.30560 0.27110
H3 0.14180 0.39070 -0.02360
H4 -0.03330 -0.29520 0.33130
Inversion-related Atoms
C1' -0.08232 -0.01856 -0.32836

C2' -0.11295 -0.16384 -0.22289
C3' -0.04799 -0.10518 -0.03714
C4' -0.07656 -0.25183 0.07582
C5' 0.01320 0.19021 -0.25460
H1' -0.12420 -0.05890 -0.45540
H2' -0.17870 -0.30560 -0.27110
H3' -0.14180 -0.39070 0.02360
H4' 0.03330 0.29520 -0.33130
results were obtained experimentally and that errors may occur either in the
diffraction experiment or in the interpretation of results. The following discussion
focuses on what to look for in a crystallographic paper and how to decide whether
or not the reported results are correct.
The two quantities that should be examined to determine the accuracy of
a crystal structure are the R values and the estimated standard deviations (esd's)
of derived parameters. The conventional R value is given by equation (17-8).
R = Y |F, - F,||Y F (17-8)
where F. are the observed structure factors derived from the measured reflection
intensities and F, are the structure factors calculated for the refined model. From
equation (17-8) it would appear that the lower the R value, the better the structure
since the observed and calculated F's will be in better agreement. Although this
is generally true, the R value may be artificially lowered using a number of
techniques. The most common way of reducing the R value of a structure is to
omit weak reflections. This is an accepted practice by most crystallographers
since all crystals exhibit weak reflections at higher diffraction angles and a cutoff
intensity (I or F) should be included in the paper. An intensity cutoff of I > 3a(I)
means that only reflections with intensities greater than 3 times their measured
deviation (sigma) will be used in calculations (3a(I) corresponds to 6a(F) since I
is proportional to F2 ). One should regard as suspect papers that report a data
cutoff greater than the 3a(I) or 6a(F) threshold. In addition to the R (or RF
values), one should be aware of the weighted R (also listed as wR or RWF). Weighted
R values are based not only on F values but on the sigma of F as well. Reflections
with higher aF will be downweighted from those with lower aF values. A typical
weighting scheme is given by equation (17-9).
w 1= a 2 (F) + gF 2 (17_9)
where g is a small value (0.01 to 0.001). The resulting weighted R value is shown
in equation (17-10).
W1/21F -- F|
RwF w 1/2 FJ (17-10)
Since the weighted R value is based on more information (aF), it is the better
value on which to base the accuracy of the structure determination. If it is reported
that a structure was refined using "unit weights" (all reflections weighted equally),
this structure should be reviewed carefully since a proper weighting scheme will
usually lead to lower esd's for derived parameters.
The best criterion on which to judge the accuracy of a crystal structure
determination is found in the derived parameters and their esd's. If interatomic
distances and angles fall within expected ranges and their esd's are small, then
it is reasonable to assume that the structural study was done properly. Omitting
too many weak reflections will lead to increased esd's so this should be avoided.
Failing to measure enough data will also result in higher esd's since the
data-to-parameter ratio will be low. A data-to-parameter ratio of at least 8 to
10:1 should be considered a lower limit for a well-defined structure, although
this is not always achievable.
Since all structural reports should include a plot of the molecule, the reliability
of derived parameters may often be determined simply by looking at the picture.
The molecular plot should be a thermal ellipsoid plot such as that shown in Fig.
17-19 1' rather than ball and stick plots as shown in Figs. 17-18(A) and 17-18(B).
The thermal motion of an atom is represented by an ellipsoid within which the
electrons of the atom have a probability of being. This probability, which is
usually 50%, should be noted on the plot. The molecule in Fig. 17-19 shows a
high degree of thermal motion for the carbon atoms (particularly the methyl
carbons) of the pentamethylcyclopentadienyl ligands even though the data set
was collected at 173 K. Since these ligands can rotate freely, it is understandable
that the carbon atoms show this high degree of libration. If an atom is disordered
(more than one possible position for an atom-static disorder), a large thermal
motion may either alert one to this possibility or may serve to "smear out" the
disordered components.
A few ions that are often disordered are Et 4N*, CIO 4 -, BF 4 and PF,-.
Et 4N' usually exhibits static disorder between the alpha carbon atoms of the
ethyl groups, whereas C1O 4 , BF 4 -, and PF,- show spherical disorder due to
their spherical nature. Figure 17-20 shows a disordered Et 4N* ion.( 12) Whenever
possible, these ions should be avoided when attempting to grow single crystals
for a diffraction study since it may not be possible to completely resolve the
resulting disorder problems. Some alternative counter ions that can be employed,
and that are usually ordered, are triflate, B(CH 5 )4 -, ((CH,) 3 P)2 N*, and
(C 6 H5 )4 As-.
FIGURE 17-19 Thermal ellipsoid plot of [(C5Me)2Sm]2(u-n2 n2-PhC 4Ph) - 2C6H5Me.
As the result of the co-crystallization of two closely related compounds, a

crystal will not be pure. Often the contaminant will be present at only a few
mole percent. This can give rise to another kind of disorder: compositional
disorder. An example is found in the structure of cis,fac-(R 3 P)3Mo(=O)Cl2 ,
which is now known to contain small quantities offac-(R3 P)3MoCl3 . An earlier
report of "anomolously" long Mo-O bond distances in the former complex
has been reinterpreted as the superimposition of the Mo-O bond with a
3
previously unrecognized small percentage of longer Mo-Cl bonds.
17-21 CRYSTALLOGRAPHIC DATA
FIGURE 17-20 Disorder of The advent of fast computers and automated data collection routines has resulted
the ethyl groups in Et4N+. in an ever-increasing amount of structural data being generated and published.
Fortunately, there are a number of computer databases and collections of printed
material that one can access to utilize this structural information.
The largest compilation of structural data is the Cambridge Structural
Database (CSD) operated by Cambridge University (England). All published or
directly deposited organic and organometallic complexes characterized by either
x-ray or neutron diffraction are included. The CSD is far more than a reference
database. Included are atomic coordinates, unit cell parameters, space groups,
R values, and esd's. Cambridge has also developed the computer programs
necessary to access the database. A graphical user interface allows one easy access
to all structural information as well as the ability to generate molecular geometries
and plots. The database contained over 82,000 entries in 1989 with 8000 to 9000
new entries yearly (Table 17-1).
The Protein Data Bank of the Brookhaven National Laboratory contains
atomic coordinates for protein and nucleic acid structures. The Inorganic Crystal
Structure Data Base is composed of all compounds lacking carbon-carbon or
carbon-hydrogen bonds but does include metal carbides. (See also reference 4.)
Exercises 711
1. The Newsletter of the Cambridge Structural Database System, Cambridge Crystallo- REFERENCES
graphic Data Centre, Cambridge, U.K., April 1990.
2. H. Hope, and B. G. Nichols, Acta Cryst., B37, 158-161 (1981).
CITED
3. G. H. Stout, and L. H. Jensen, "X-Ray Structure Determination. A Practical Guide,"
2nd ed., Wiley, New York, 1989.
4. J. P. Glusker and K. N. Trueblood, "Crystal Structure Analysis. A Primer," Oxford
University Press, Oxford, U.K., 1985.
5. D. Brodalla, D. Mootz, R. Boese, and W. Osswald, J. Appl. Cryst., 18, 316-319 (1985).
6. M. C. Etter, and P. W. Baures, J. Amer. Chem. Soc., 110, 639-640 (1988).
7. a. "International Tables for X-Ray Crystallography," Vol. III, eds. C. H. Macgillavry
and G. D. Rieck, Kynoch Press, Birmingham, England, 1962.
b. "International Tables for X-Ray Crystallography," Vol. I, eds. N. F. M. Henry and
K. Lonsdale, Kynoch Press, Birmingham, England, 1952.
8. H. Hope, University of California, Davis. Personal communication to J. Ziller, 1987.
9. R. E. Marsh, Acta. Cryst., B42, 193-198 (1986).
10. R. E. Marsh, and F. H. Herbstein, Acta. Cryst., B44, 77-88 (1988).
11. W. J. Evans, R. A. Keyer, and J. W. Ziller, Organometallics, 9, 2628-2631 (1990).
12. M. R. Churchill, and S. W.-Y. Chang, Inorg. Chem., 13, 2413-2419 (1974).
13. K. Yoon, G. Parkin, A. L. Rheingold, J. Am. Chem. Soc., 113, 1437-1439 (1991).
1. A crystal of [InIMes2] 2 is monoclinic with reciprocal lattice parameters a* = 0.038930, EXERCISES

b* = 0.058389, c* = 0.127765 and #* 75.298. Calculate the direct cell parameters and
unit-cell volume. (Mes is mesityl).
2. The density of the above complex is 1.792 mg/m 3 . The number of molecules in the unit
cell may be calculated using: n = [(density)(V)]/[(Avogadro constant)(MW)]. What is
n (Z value) for the complex?
3. The systematic absences observed for the above complex are hkl for h + k = 2n + I
and hOl for 1 = 2n + 1. What are the two possible monoclinic space groups? Given the
calculated Z value, is it possible for the molecule to lie on a general position in both
space groups? What point symmetry is possible in one of these space groups? (See
International Tables.)
4. Determine the space group(s) consistent with the following systematic absences and
crystal class.
a. Monoclinic: OkO for k = 2n + 1; hOl for h + I = 2n + 1.
b. Orthorhombic: hkl for h + k, k + l; Okl for k + l = 4n + 1; h0l for n + l = 4n + 1;

hk0 for h + k = 4n + 1
c. Orthorhombic: hkl for h + k + 1 = 2n + 1; Okl for k = 2n + 1; hOl for h = 2n + 1
d. Orthorhombic: Okl for k = 2n + 1; hOl for = 2n + 1; hk0 for h = 2n + 1
e. Same as for (d) but also hkl for h + k + l = 2n + I
f. Triclinic: hkl for h + k + l = 2n + 1

3
5. What is the absorption coefficient (p) for (MeC,)3 Sm given the density is 1.375 mg/m ?
6. What is the number of accessible reflections for the complex in question (1) for both
Mo and Cu radiation and how many octets of data will need to be collected? Why
may the resulting number of reflections appear to be less than required? How many
reflections will there be for Mo radiation with a 2-theta limit of 55 degrees?
7. If a molecule has a chiral center and is known to be resolved, what is a specific reason
why it may not crystallize in space group P21/c?
8. What will be the maximum and minimum absorption effects on the reflections from
the crystal in problem 5 if the crystal measurements are 0.10 x 0.40 x 0.80 mm? Should
a correction be made? Should this crystal be used in the X-ray diffraction study?
APPENDIX
Character Tables for

Chemically Important A
Symmetry Groups
1. THE NONAXIAL GROUPS
C, E
A 1
C, E ah C,
A' 1 1 x,y,R 2 2
x ,y , A,
zz22,xy
A" 1 1 z, Rx, R, yz, xz A
2. THE C,, GROUPS
C2 E C2
A 1 1 z, Rz x 2 , y 2 ,z 2 , xy
B 1 -1 x, y, Rx, R, yz, xz
C3 E C3 C3 2 C = exp (27zi/3)
2 2
A 1 1 1 z, Rz x2 + y2, z
E 1 *(x, y)(Rx, R,) (x2 _ Y2 xy)(yz, xz)
(x, y)(Rx, R,) (yz, xz)

714 Appendix A Character Tables for Chemically Important Symmetry Groups
2. THE C,, GROUPS (continued)
E C5 C 52 C5 3 C5 4 E = exp (2ri/5)
1 1 1 1 1 z, Rz x2 + Y2, z2
2* F
2
9 2* 2 (x, y)(Rx, R,) (yz, xz)
1 E*
2
g 2* F2
* 2 (x 2 _ ,Xy)
C6 |E C6 C, C2 C3
2 C 65 S=exp (2nri/6)
-~ I.
X2 +Y2, 2
1 1 1 1 1 z, Rz x2+y2,z2
1 1 -1 1 - l
(X, y)
F -s* -1
* 8
(Rx, R,) (xz, yz)
* -F -1
F* -F 1 F* -F
X2 _ 2,
-F -F* 1 F -EJ
C7 E C7 C 72 C 73 C7 4 C7 5 C7 6 F exp (2ri 7)
A 1 1 1 1 1 1 1
2
1 2 2 3 3 *
E *
E2 2 3
8
* * 3 2
2 3 3
F * F*
(X2 _ 2
E2 2 2*
* 3 3F 2 , Xy)
3 2
C F.* eF
2
2* 3
E3 9 eF,** 2
* 3
3 5 7
C, E C8 C4 C2 C4 3 C8 C8 C8 E = eXp (27ri 8)
A 1 1 1 1 1 1 1 1 z, Rz x2 + Y2, z2
B 1 1 1 1 1 -1 -1 -1
-EF -EF e (x, y) (xz, yz)
(Rx, R,
Si -1 1
1 -i i -i
E2 _ 1 (X2 _ ,Xy)
E3
3. THE D. GROUPS
D2 E C3(z) C2 (y) C2 (X) I

A 1 1 1 1 x 2+y2,z2
B, 1 1 -1 -1 z,Rz xy
B2 1 -1 1 -1 y, R, xz
B3 1 -1 -1 1 x,Rx yz
D3 E 2C 3 3C 2
Al 1 1 1 x2+Y2,z2
A2 1 1 -1 z,Rz
E 2 -1 0 (x, y)(Rx, R,) (x 2 _ y 2, xy)(xz, yz)
Appendix A Character Tables for Chemically Important Symmetry Groups 715
D, E 2C4 C2 (= C4 2 ) 2C 2 ' 2C 2 "

x 2 + Y,2 z 2
z, Rz
2
X2
xy
(x, y)(R., R,) (xz, yz)
3. THE D, GROUPS (continued)
D5 | E 2C5 2 5C 2 |
x 2 + y 2, z 2
z, Rz
2 cos 72' 2 cos 144' (x, y)(Rx, R,) (xz, yz)
2 cos 144' 2 cos 72' (x 2 y 2 xy)
D6 E 2C6 2C, C2 3C 2 ' 3C 2 "

X2+Y2,
x2+y2,zz
2
(x, y)( R;, R,) (xz, yz)

2
(x _ y2)
4. THE Cn, GROUPS
C2 " | E C2 an(xz) a6'(yz) | 2 2 2

A1 z X 2, 2, z2
A2 Rz xy
B1 x,R, xz
B2 y, Rx yz
"For a planar molecule the x-axis is taken perpendicular to the

plan e.
C E 2C, 3,,
X2+Y2, 2
z x y2,z2
Rz
(x, y)( R., R,) (x 2 _ y2 , xy)(xz, yz)
E 2C 4 C2 2a, 2
aGd
2
z x + y 2, z 2
Rz
x2 _ 2
x y
(x, y)(Rx, R,) (xz, yz)
"If the C4 , molecule contains a square array of atoms, the a, planes should pass
through the larger number of atoms of the square array or should intersect the largest
possible number of bonds.
E 2C, 2C5 2 5u,

A1 1 1 1 1 X2 +Y2z2
x2+y2,z2
A2 1 1 1 -1
E, 2 2 cos 72 2 cos 144 0 (xz, yz)
E2 2 2 cos 144' 2 cos 72 0 (x 2 _ ,y)
Co6 |
E 2C6 2C, C2 3a 3a,
I
I
Al 1 1 1 1 1 1 x 2+y2, z2
A2 1 1 1 1 -1 -1
B1 1 -1 1 -1 1 -1
B2 1 -1 1 -1 -1 1
El 2 1 1 -2 0 0 (x, y)(R., R,) (xz, yz)
E2 2 1 -1 2 0 0 (x 2 _ y2 , XY)
5. THE C,,h GROUPS
C2 I E C2 i 6,
A, 1 1 1 1 R, x 2 ,y 2,z 2,XY
B, 1 1 1 -1 R xz, yz
A. 1 1 1 1 z
B. 1 1 1 1 x,y
E C3 C 32 Uo S3 s3 e = exp (27ri/3)
I
I
1 1 1
r
1 1 1 R, x2 + y2, z2
1~E* 1 8
1 1 1 -1 -1 -1
r
z
1~e* -1 -e ej (R,, R,) (xz, yz)
E C4 C2 C43 i S43 eh S4
Ag 1 1 1 1 1 1 1 1 X2+Y2, 2
x2+y2,z2
Bg 1 -1 1 -1 1 -1 1 -1 x2 y2
1 i -1 -i 1 i -1 -
E, (xz, yz)
A, 1 1 1 1 -1 -1 -1 -1
B, 1 -1 1 -1 -1 1 -1 1
E".
1 -i -1 i -1 -i 1 -i
Csh E C5 C5 2 C5 3 C 54 Uh S5 S5 S53 S59 8= exp (2mi/5)
A' 1 1 1 1 1 1 1 1 1 1 R, x2 + Y2, z2
1 E E2 E2* 8* 1 9 2 E2* 8*
E 1 * 2* E 2** E 82 E y)
E2' 1 82* E* 2 2 2
82 8* 82 1 82* 2(8*y
A" 1 1 1 1 1 1 1 1 -1 -1 z
1 E 82* 82 8 -1 -8* -E 2* *
E1" 2 * 1
_2* - _ 2 -82 _E (R,, R,) (xz, yz)
E E2* _1 _ 2 _E* 8 _ 2*
E" E*
2" 2* 8 8 2 1 -E2* -E * 82
2 5 5
C,, E C, C3 C2 C3 C. i S3 S65 Ch S, S, =exp (2ri/6)
A, 1 1 1 1 1 1 1 1 1 I 1 1 R: X2+ Y2,Z2
B, 1 1 1 - 1 1 1 - 1 1 1 1
Ei, , _1 E _(R, R,) (xz, yz)
18-* -s 1 -E* - 1 -E - 1 8 -E
E2g -F8, v* X Y2X)
A, 1 1 1 1 1 1 1 1 1 1 1 1 z
B, 1 -1 1 -1 1 -1 -1 1 -1 1 -1 1
El.
E2 , 8
l -E*E -
- -1I -E*-8* - 8 -1
-1 -8*
F* 8 -lY*)1 8* - 1 (X
(1 - * -1 8 * -1 8 *j
6. THE Dflh GROUPS
D2h E C 2 (z) C2(y) C2 (x) i a(xy) U(xz) o(yz)

A, 1 1 1 1 1 1 2
B 1, 1 1 -1 -1 1 1 -1 -1 Rg xy
B2 g 1 -1 1 1 1 -1 1 1 xz
B 39 1 1 1 1 1 1 1 1 Rx yz
A, 1 1 1 1 1 1 -1 1
B 1g 1 1 1 1 1 1 1 1 z
B2. 1 1 1 -1 -1 1 -1 1 y
B 3u 1 1 1 1 1 1 1 1 x
D 3h E 2C 3 3C2 ah 2S3 3a,,

A,' 1 1 1 1 1 1 x2+y2,z2
A2 ' 1 1 -1 1 1 -1 R2
E' 2 -1 0 2 -1 0 (x,y) (x 2 y 2,XY)
A1 " 1 1 1 -1 -1 -1
A2" 1 1 -1 -1 1 1 z
E" 2 -1 0 -2 1 0 (R, R,) (xz. yz)
2
D4" E 2C 4 C2 2C 2 ' 2C 2 " i 2S4 Ch 2a, a,
Ajg 1 1 1 1 1 1 1 1 1 1 x2 +y2, 2
x2+y2,z2
A 2g 1 1 1 -1 - 1 1 1 1 1 R.
x2 _Y2
Blg 1 1 1 1 1 1 1 1 1 -1
B2g 1 -I 1 -1 I I 1 1 1 1 xy
Eg 2 0 -2 0 0 2 0 2 0 0 (R,, R,) (xz, yz)
Aiu 1 1 1 1 1 -1 -1 -1 - -1
A2. 1 1 1 -1 1 --1 -1 --1 1 1 z
Bia 1 -1 1 1 -I -1 1 -1 -1 1
1 -1 1 -1 1 -1 1 -- l 1 -1
Eu 2 0 -2 0 0 -2 0 2 0 0 (x, y)
"a, passes through the atoms and ad bisects the bond angles.
Ds I E 2C 5 2 5C 2 ah 2S5
2 2 2
Ai 1 1 1 1 1 1 x2+y2,z
A2 ' 1 1 1 - 1 1 1 R,
El'l 2 2 cos 72 2 cos 144 0 2 2 cos 72" 2 cos 144' (x, y)
E2' 2 2 cos 144 2 cos 72 0 2 2 cos 144' 2 cos 72' (x 2 _ y2' Xy)
Al" 1 1 1 1 1 1
A 2" 1 1 1 -1 1 1 z
El," 2 2 cos 72 2 cos 144 0 -2 -2 cos 72' 2 cos 144 (Rx, R,) (xz, yz)
E2" 2 2 cos 144 2 cos 72 0 -2 -2 cos 144' 2 cos 72'
Don I E 2C6 2C3 C2 3C2' 3C2" i 2S3 2S6 a, 3as 3a,
Dh
~I. E 2C6 2C3 C2 3C2 ' 2S3 2S6 c,, 3q4 3o~, t
Ai, 6
1 1 1 1 1 1 1 1 1 1 1 1 x 2+y2, z2
A2 y
1 1- 1 -1 1 1 1 -1 1 -1 1 -
Big
B 2g 1 1 1 -2 1 1 -1 1
Elg 2 1 1 2 0 0 2 1 -1 -2 0 0 (R,, R,) (xz, yz)
E2g 2 1 -1 2 0 0 2 -1 -1 2 0 0 (x 2 _ y 2, xy)
Alu 1 -1 -1 -1
A2u 1 -1 -1 -1 z
Blu 1 -1 1 -1
1 - 1 1
B2u 1 -1 1 - 1
Em 2 1 1 2 0 0 -2 1 1 2 0 0 (x, y)
l 2 -1 -1 2 0
E2. 0 -2 1 1 2 0 0
6. THE Dflh GROUPS (continued)

Deh | E 2C 8 2C 8 3 2C 4 C2 4C 2 ' 4C 2 " i 2S, 2S, 3 2S 4 ah g 4ul,
-I. i I
A lg 1 1 1 1 1 1 1 1 1 1 x2+y2, 2
x2+y2,z2
A2g 1 1 1 1 1 1 1 1 -1 -1 R,
B1g -1 -1 1 1 1 -1 1 1 1 -1
B2g -1 -1 1 1 -1 - 1 1 - 1 1
Eig 2 - 2 0 -2 2 -2 0 -2 0 0 (R,, R,) (xz, yz)
E2g 0 0 -2 2 0 0 -2 2 0 0 (x 2 _ ,Xy)
E3, - 2 2 0 -2 -2 /2 0 -2 0 0
Alu 1 1 1 1 -l - I - - -1 -1
A2u 1 1 1 - 1 -I - - -1 -1 1
Blu -1 - 1 1 1
1 1 - 1 - 1 - 1 1
B2. -1 -1 1 1
Eil 2 - '2 0 -2 -2 /2 0 2 0 0 (x, y)
E2u 0 0 -2 2 0 0 2 -2 0 0
E3. 2 2 0 -2 2 -/2 0 2 0 0
7. THE Dd GROUPS
E 2S4 C2 2C 2 ' 2
ad
1 1 1 1 1
1 1 1 -1 -1
X2_Y2
1 -1 1 1 -1
1 -1 1 -1 1 x y
1 0 -2 0 0 (xz, yz)
E 2C, 3C2 i 2S6 3

Ud
X2+Y2, 2
1 1 1 1 1 1 x2+y2,z2
2 1 1 1 1 1 Rz
2 1 0 2 1 0 (Rx, R,) (x 2 _ y 2, Xy)
(xz, yz)
1 1 1 -1 -1 -1
1 1 -1 -1 -1 1 z
2 -1 0 -2 1 0 (x,y)
D4d E 2S, 2C4 2S, 3 C2 4C 2 ' 4aI

A1 1 1 1 1 1 1 x 2+y2,z2
A2 1 1 1 1 -1 -1 R2
Bi -1 1 -1 1 1 -1
B2 -1 1 -1 1 -1 1 z
Eil 2 0 - 2 -2 0 0 (x, y)
E2 0 -2 0 2 0 0 (x 2
_2 XY)
E3 2 0 -2 2 0 0 (R,,R,) I(xz, yz)
E 2C, 2C, 2
5C2 i 2S10 3 2S10 5
0d
l11 1 1 1 1 1
R.
7
2 2 cos 2 2 cos 144 0 2 2 cos 72 2 cos 144 0 (Rx, R,) (xz, yz)
2 2 cos 144 2 cos 72 0 2 2 cos 144 2 cos 72 0 (x 2 _ ,Xy)
1 I 1 -1 -1 - 1 -1 1 z
2 2 cos 72 2 cos 144 0 -2 - 2 cos 72 -2 cos 144 0 (x, y)
2 2 cos 144 2 cos 72 0 -2 -2 cos 144 -2 cos 72 0
7. THE Dfd GROUPS (continued)

D6d I E 2S12 2C6 2S, 2C 3
2S 12 5 C2 6C 2 ' 6
a, 4 4
A7 1 1 1 1 1 1 1 1 X2+Y2z2
x2+y2,z2
A2 1 1 1 1 1 1 -1 -1
B12 -1 1 -1 1 -1 1 1 -1
B2 --1 1 -1 1 -1 1 -1 1 z
E1l 3 1 0 1 3 -2 0 0 (x, y)
E2 1 - 1 -2 - 1 1 2 0 0 (x 2 _ ,Xy)
E3 0 -2 0 2 0 -2 0 0
E4 1 -1 2 1 -1 2 0 0
Es5 3 1 0 -1 3 2 0 0 (R,, R,) (xz, yz)
8. THE S, GROUPS
3
S I E S, C2 S4
X2+Y2, 2
A 1 1 1 1 Rz x2+y2,z2
B 1 -1 1 -1 z x2 2
E (x, y); (R., R,) (xz, yz)
S6 E C3 C3 2 i s65 So e = exp (276/ 3)
1 1 1 1 1 1 R, x2+y2,z2
2 2 2
(Rx, R,) (X2 _ , xy);

(xz, yz)
1 1 1 -1 -1 -1 z
(X, y)
(1e 1 -E* - *
S8
A
B
E,1
E
1
1 -1
S8
1
1
C4
1
1
S8 3
-
-1
1
*
C2
1
1
1
S8 5
-1
-E
1
C4 3
-
1
1
S?
-1
1
E*1
I
RK
z
(X,y);
e =exp
x2 +y2,
(2nri/8)
z2
(1 s* -i -s -1 -s* i eJ (Rx, R,)
{1
1- -1
- * -
i
8 -1
r
8*
-
ij s
(X222 y2 , Xy)
(Xz, yZ)
9. THE CUBIC GROUPS
T E 4C, 4C 3 2 3C2 t = exp (2iU/3)
A 1 1 1 1 2+Y2+z2
1 g 1* 1 (2z2 _ 2 _ ,
E 1 X2 _ 2)
T 3 0 0 -1 (R,, R,, Rz); (x, y, z) (xy, xz, yz)
9. THE CUBIC GROUPS (continued)
T E 4C3 4C32 3C 2 i 4S 6 4S 6 5 3
-h E- exp (2ci3)
1 1 1 X2 +Y2 + 2
x2+-y2+z2
1 -1 -1
0 2
0* 1 (2z _ X2 _ y2
2
x _2)
-g
-8*
-8*
-8
-18 -l
o 0 -1 (Rx, R,, R.) (xz, yz, xy)
o 0 1 (x, y, z)
TI E 8C 3C2 6S4 6-,

1 1 1 1 X2+Y2 +z2
x2+y2+z2
1 1 -1 -1
-1 2 0 0 (2z 2 _ X2 _2
x 2 _ ,2)
T, 3 1 1 - 1 (Rx, R,, R2)
T2 3 -1 -1 1 (x, y, z) (xy, xz, yz)
0 E 6C4 3C2(= C42) 8C3 6C2
X2+Y2 +z2
(2z 2 X 2 _,
2
X _ Y2)
1 - 1 0 -1 (R,, R,, Rz); (x, y, z)
1 -1 0 1 (xy, xz, yz)
E 8C3 6C 2 6C 4 3C2 ( C42) i 6S 4 8S 6

3
a, 6
ad
A, 1 1 1 1 1 +, + ..2
A 2g 1 1 -1 -1 1 2
x2+y +
E, 2 1 0 0 2 2 0 1 2 0 (2z'z_ 2 _y2,
2
x _ 2
Tta 3 0 1 1 1 3 1 0 1 1 (R,R,,Rz)
T2 , 3 0 1 -1 1 3 1 0 1 1 (xz, yz, xy)
A , I I I 1
A 2. I I I - I
E. 2 1 0 0 2 2 0 1 2 0
T, 3 0 1 1 1 3 1 0 1 1 (x. y, z)
T'.. 3 0 1 1 - 3 1 0 1 1
10. THE GROUPS C, AND Deh FOR LINEAR MOLECULES
C- E 2Cj* .. _a_
X2+Y2, 2
A1 -+ 1 1 . . 1 z x2+y2,z2
Ai--I+ 1 1 .. -1 Rz
2 2 cos D ... 0 (x, y); (R,, R,) (xz, yz)
E2 = A 2 2 cos 20 0 (x 2 _ ,y)
E3 - D 2 2 cos 3( . 0
D, IE 2Cj* ... wav i 2S.* ... 0C2
X2 +Y2,
x2+y2,z2 2
R.
2 2 cos (D 0 2 -2cos D .. 0 (R,, R,) (xz, yz)
2 2 cos 2)D . 0 2 2 cos 20D . 0 (x 2 _Y2 xy)
1 1 .. 1 1 1 .. 1
1 1 1 1 1 . . 1
z
2 2 cos (D 0 -2 2cos ( ... 0 (x,y)

2 2 cos 20 0 -2 2 cos 20 ... 0
11. THE ICOSAHEDRAL GROUPS"

3
I, E 12C, 12C,
2
20C 3 15C2 I i 12Sjo 12S, 20S, 15-
i I 1 I I 1 1 1 1
0
1 1 X2+ . + Z
3 (I + ) 2 I{- , 5) 0 1 3 -1- , 5) '(1 + ) 0 (RR, R)

3 j1 1- )
53) 1 ++1 5 )
2~)
0
0 - 1 3 -%,
3) . 5 '(1
l 5 0
0 1I
4 1 1 1 0 4 -1 -1 1 0
5 0 0 1 1 5 0 0 1 1 (2:2 _X2 - 2
xt, vz zx)
A, 1 1 1 -1 I I- I I
T, 3 + ) (1 - 5) 0 -3 -}(1 - 5) -'(1 + 5) 0 1 (x. ,:
T2 3 - ) +(1 1(+
5) 0 - 3 -(1 + 5) -W ( 5) 0 1
G, 4 - - 1 0 -4 1 1 1 0
H. 5 0 0 1 1 -5 0 0 1 -
For the pure rotation group I, the outlined section in the upper left is the character table; the g subscripts should, of course, be dropped and (x, v, :)
assigned to the T, representation.
APPENDIX
Character Tables for

B Some Double
Groups
GROUP C3'
C3 ' E R C, CR C3 2 C3 2 R
A' F1 1 1 1 1 1 1 z,L,
E' F2 1 1 2 0 2- yL)
, ,L
E , 1 - -w - 0) 2 2 X
E1/2' F 1 1 _ 2 - 2 _2 )2
F, 1 )2 ) (0 cW
B3 / 2' F 1 - 1 1 1 -1
o = exp (iir/3)
GROUPS D 3 ' AND C,,'
D3' E R C 2R
2 C 3C22____
' 3C2____
2 'R D3 '
C3 R C3R
C3 ,' E R C C 3, R C3,'
C3 23R C 33R _____
A 1' F1 1 1 1 1 1 1 z
A2 ' F2 1 1 1 1 -1 -1 z, Lz Lz
E' F3 2 2 -1 -1 0 0 x, y, Lx, L, x, y, L, L,
E 3 /2' (A1 ') F4 1 -1 -1 1

((A2 ') Fs 1 -1 -1 1
E 1/ 2 (E') F, 2 -2 1 -1 0 0
Appendix B Character Tables for Some Double Groups 725
GROUPS C4' AND S4'

3
C4 ' E R C4 C4 R C4 2 C4 2 R C4 C 4 3R C 4'
S4' E R S4 S4 R C2 R C2 S4 3 S4 3 R S4'
A' F1 1 1 1 1 1 1 1 1 z L
B' F2 1 1 -1 -1 1 1 1 -1 L z
E' 3 1 1 i i -1 1 -i i iy Lx+iy
F4 1 1 -i -i -1 -1 i i L iL3, L iL,
_ 3
E1 2 ' F53 10
IF2 _ _ J 3 3 ( t
3 3
I 03 i _ -(t
E ' F7 1 3
3/ 2 8 1 _ 1 _ _i
wo= exp (in/4)
GROUP T'
T' E R 4C32 4C 3 2 R
3 3C
C 2 2R
4C3 4C 3 R
A' F1 1 1 1 1 1 1 1
E 2 ( 1 1 11 ) ( 2 )2
E 3 1 1 02 (2 2
T' F4 3 3 -1 0 0 0 0 xy, z, Lx, L,, L2
E 1 2' F5 2 -2 0 1 -1 1 -1
G3/2 , 2 -2 0 02 -a w02 _,_2
3/2 1 2 -- 2 0 2 _ 02 _(0
ow= exp (2nri/3)
GROUPS O' AND Td
0' E R 4C,
4C 3 2 R
4C 2
4C 33 R
3C422
3C4 R
3C
3C434 R
3C, 3
3CR
3C2'
3C2 'R
o0
4C,2 4C 3 2 3C22 3S, 3
3S4 6
6
0d T'
Td' E R 4C R 4C 3 R 3C 44 R 3S 4 3R 3SR dR d
A,' F, 1 1 1 1 1 1 1 1
A2 ' F2 1 1 1 1 1 - 1 1
E' F3 2 2 -1 -1 2 0 0 0
T1' [4 3 3 0 0 -1 1 1 -1 xy z,L.,,L,,L Lx,,L,,L
T' F, 3 3 0 0 -1 -1 -1 1 xyz
(El') E 1 2' F6 2 -2 1 -1 0 2 2 0
(E2') E5 2' [, 2 -2 1 -1 0 - 2 0
(G') G 3 2' F8 4 -4 -1 1 0 0 0 0
726 AppendIx B Character Tables for Some Double Groups
GROUP D4'
D' E R C 3 C2 2C 2 ' 2C 2 "
C4 3R 4 C4 R C2 R 2C2'R 2C 2 "R
A ' F1 I 1 1 1 1 1 1
A 2'FE 1 1 1 1 1 1
B'F 3 I 1 1 1 1 1 1
B2' F 4 I 1 1 1 1 1 1
E ' F 5 2 2 0 0 -2 0 0
E2 ' F 6 2 -2 2 2 0 0 0
E3 ' F 7 2 -2 2- 2 0 0 0
APPENDIX
Normal Vibration
Modes for Common
Structures
C
x
v1(A1) d2(A
l) 13
M(E)s V4.(E)
Pyramidal XY3 Molecules.
(Y2
vI(A1) v 2 (E) P3(T2)
Tetrahedral XY 4 Molecules.
Y3 +
x-
+ +a"
y1 2
1(A'1) P2(A"2) P(E') 4(E)
Planar XY 3 Molecules.
727
728 Appendix C Normal Vibration Modes for Common Structures
1(A1) 3 (AI)
F6(E)
C3 , ZXY3 Molecules.
Y3 (Fl.)
"4(11J 16(T2u)
Octahedral XY, Molecules.

Appendix C Normal Vibration Modes for Common Structures 729
Y1(AIg)
V6 (Eu)
Y M(El)
Square-Planar XY4 Molecules.
v4 (B2 )
Planar ZXY 2 Molecules.

730 Appendix C Normal Vibration Modes for Common Structures
-Y
Y. ......
Y
1(A',) Y2(A'I) P3(A~2)
P7 (E') v 8 (E")
Trigonal Bipyramidal XY, Molecules.
X
Y
v1(AIg) 112(A Ig) Y3(A9) v4(A1.)
us(A 2) "rs(A 2.) Y7(E.)
v 12(Eg)
Ethane-Type X2Y6 Molecules.

Appendix C Normal Vibration Modes for Common Structures 731
XI X2
"'0
Y1/ y Y2
P.(XY) P(XX) 6,(YXX)
Planar (C2,) A 1 (p) A1 (p) A I(p)

Non-planar (C 2) A(p) A(p) A(p)
Yas(XY) bas(YXX) P1 (XY)
B2(dp) B2(dp) A2 (dp)
B(dp) B(dp) A (p)
Nonlinear X 2 Y 2 Molecules (p: Polarized; dp: Depolarized).

APPENDIX
Tanabe and Sugano

D Diagrams for Oh Fields*
DO/B
Energy diagram for the configuration d2.
* These are complete energy diagrams for the configurationsindicated, reproduced from Y. Tanabe
and S. Sugano, J. Phys. Soc. Japan, 9, 753, 766 (1954).
y refers to the ratio of the Racah parameters C/B. Heavy lines perpendicular to the Dq/B axis in
4
d , d', d', and d' indicate transitions from weak to strong fields. The calculation of and the assumptions
inherent in the calculations are contained in the original paper. The diagrams do not apply to any
particular complex but give a qualitative indication of the energies of the various states as a function
of Dq/B.
Dq/B Dq/B
Energy diagram for the configuration d3. Energy diagram for the configuration d4.
Dq/B Dq /B
5
Energy diagram for the configuration d . Energy diagram for the configuration d6.
734 Appendix D Tanabe and Sugano Diagrams for 0
, Fields
4'A(d3d'
. ,2 1
DI/B Dq /8
Energy diagram for the configuration d7 . Energy diagram for the configuration d 8 .
APPENDIX
Calculation of
A (10Dq) and # for
Oh Ni" and Td Co"
Complexes
CALCULATION OF A AND p FOR OCTAHEDRAL
Ni2 + COMPLEXES
The data in Table 10-6 for the Ni[(CH 3 )2 SO] 6 (Cl0 4 ) 2 complex will be employed
3
to illustrate the calculation of A, #l, and the frequency for the 3A 2,g -+ T(F)
band. The value for A, or 1ODq, is obtained directly from the lowest energy
transition, 3A 2 -+ 3 T2., which occurs at 7728 cm-'. Equation (10-12),
[6Dqp - 16(Dq) 2] + [ -6Dq - p]E + E2 = 0,
is employed to calculate the experimental 3P energy value, that is, p of equation

(10-12). The quantity p is equal to 15B for nickel(II), where B is a Racah parameter.
Racah parameters indicate the magnitude of the interelectronic repulsion between
various levels in the gaseous ion. The quantity B is a constant that enables one
to express the energy difference between the levels of highest spin multiplicity in
terms of some integer, n, times B; that is, nB. Both n and B vary for different
3 3
ions; in the case of Ni 2 +, the energy difference between F and P is 15B. The
same term adjusted for the complex is 15B'. To use equation (10-12) it is necessary
3
to employ the energy values for the T,,(P) state. This is the energy observed
3 3
for the Tg(P) transition (24,038 cm- ) plus the energy of the A2, level, because
E observed for transition (24,038 cm-i) = energy of 'Tg(P) - energy of 3 A2
so
energy of 3 Tg(P)= E for transition + E of 3
22g
Thus, combining this with equation (10-11) (E of

3
22 = -12Dq; note that
Dq = 7728/10;~ 773), we obtain
E of 3 Tt(P) = 24,038 cm 12Dq = 14,762 cm
The value of E = 14,762 cm-' is employed in equation (10-12) along with

Dq = 773 to yield a value
p = 13,818 cm-' = 15B'

0
736 Appendix E Calculation of A (10Dq) and # for Ni" and Td Co" Complexes
The gaseous ion E(3 P) value for Ni 2 + is 15B = 15,840 cm-1 and # is [equation
(10-14) or (10-13)]:
13,818 15B'
- =- - 0.872
15,840 15B
or
15,840 - 13,818
# = 1584 x 100 =12.8 %
15,840
To calculate the energies for the 3 Tg(F) and 3 Tig(P) states, the values p = 13,818
cm - I and Dq = 773 cm 'are substituted into equation (10-12) and the equation
is solved for E. Two roots, E = 14,762 cm-' and E = 3,694 cm-', are obtained.
Since the transition is 3A 2 g 3
3- T(F) or 3
Tg(P), the absorption bands will
correspond to the differences
E[ 3 Tg(F)] - E[ 3A2 ] and E[ 3 T'g(P)] - E[ 3 A 2,]
or
[3 A2g 3 Tl(F)] = 3,694 - [- 12(773)] = 12,970 cm 1
and
[3A2,-
3
Tg(P)] 14,762 - [ 12(773)] = 24,038 cm
The agreement of the calculated and experimental values for the 12,970 cm-1
band supports the # and Dq values reported above.
CALCULATION OF A AND P FOR T, C02+ COMPLEXES
In a field of tetrahedral symmetry, the 4F ground state of Co 2 + is split into "A ,

2
4T2 , and 4T,(F). The transitions "A2 -*T 2 , "A2 -* 4T,(F), and 4A 2 -- 4T,(P) are
designated as vi, v2 , and v 3 , respectively. The following relationships are used
to calculate A and #l:
v1= A (E-1)
v2 = 1.5A + 7.5B' - Q (E-2)
v3 = 1.5A + 7.5B' + Q (E-3)
1
- 2 [(0.6A - 15B') 2 + 0.64A 2 ] 1 2 (E-4)
where B' is the effective value of the Racah interelectronic repulsion term in the
complex. To repeat equations (10-14) and (10-13),
B'
_ complex
Bfree ion
Appendix E Calculation of A (10Dq) and # for
0
h Ni" and Td Co" Complexes 737
or
1 Bfeio
- -- B' complex
X 100 (E-5)
Bfree ion
To demonstrate the calculation, let us consider the spectrum of tetrahedral

Co(TMG)4 2 + (where TMG is tetramethylguanidine).(') The band assigned to v3
is a doublet with maxima at 530 mp (18,867 cm-'), e = 204, and 590 mp (16,949
cm-1), - = 269. The near infrared spectrum yields v2 as a triplet: 1204 mp (8306
cm-'), r = 91.5; 1320 mp (7576 cm-'), E = 85.0; and 1540 mp (6494 cm-'),
2
e = 23.5. The T, states are split by spin-orbit coupling to the following extent( ):
4 4
-%/4, + 6/42, and + "/42. The energy of A 2 - T,(F) is obtained by averaging
the three peaks for the v2 band, using the above weighting factors.
9
- (6494) = 14,612
4
6
- (7576) = 11,364
4
- (8306) = 31,148
4
30
Totals - 57,124
4
The average energy of v2 is thus 57,124 +30/4 = 7617 cm~-. The energy of the
transition from 4A 2 to 4 T 1(P) (i.e., v 3 ) is obtained by averaging the two peaks to
produce 17,908 cm-'. The series of equations (E-1) to (E-4) are now solved to
obtain A and fl. Adding equations (E-2) and (E-3) produces:
v2 + v3 - 15B'
3
2
and substituting v2 and v 3 for Co(TMG)4 + produces A = 5(1702 cm-' - B')
Subtracting equation (E-2) from (E-3) produces
1
Q= - (v
2 3
- v2 ) = 5146 cm-
Squaring both sides of equation (E-4) and rearranging produces
4Q 2 = A2 - 18B'A + 225(B')2 (E-6)
Substituting Q= 5146 cm-' and A = 5(1702 cm-' - B') into equation (E-6)
yields an equation that can be solved from B'. One root is 821 cm-1, and the
other root is negative. When the positive root is substituted into A = 5(1702
cm-' - B'), the value of A = 4405 cm-' is obtained; #lis evaluated from equation
(E-5).
738 Appendix E Calculation of A (10Dq) and # for Oh Ni" and Td Go" Complexes
REFERENCES 1. R. S. Drago and R. L. Longhi, Inorg. Chem., 4, 11 (1965).

2. R. Stahl-Broda and W. Low, Phys. Rev., 113, 775 (1959).
CITED
APPENDIX
Conversion of
Chemical Shift F
Data
The chemical shifts, J (in ppm), relative to o = 0 for tetramethylsilane for some
compounds often employed as external standards are: cyclohexane, - 1.6;
dioxane, -3.8; H20, -5.2; CH 2 Cl2 , -5.8; C 6 6 , -6.9; CHCl3, -7.7; and
H2 SO 4 (sp. gr. 1.857), -11.6 ppm. (The larger negative value indicates less
shielding.) These shifts are obtained on the pure liquids relative to an external
standard. As a result, they can be employed to convert data and allow comparison
of results between the various materials as external standards. To convert 6
obtained toward C6 H6 as a reference to Si(CH 3 )4 (external standard), subtract
6.9 ppm from the C6 H value. One should check to be sure that the sign convention
for A is that described for protons (a, -- R).
The conversion of results obtained relative to an external standard to an
internal standard is not quite as straightforward. If chemical shift values at infinite
dilution in CC14 are converted to Si(CH 3 )4 as a reference by using the above
data, T values do not result. This procedure converts the data to the reference
pure Si(CH 3 )4 . The difference in ( for Si(CH 3 )4 in the pure liquid and at infinite
dilution in CCl 4 is about 0.4 ppm. The pure liquid is more shielded.
For fluorine shifts, F 2 = 0 ppm is often taken as the standard. Shifts, in
ppm, for other liquids relative to F2 are SF6 , 375.6; CFCI3 , 414.3; 3 Cl, 454.2;
CF4, 491.0; CF 3COOH, 507.6; C 6HF, 543.2; SiF 4 , 598.9; and HF, 625, where
the positive value indicates a more highly shielded fluorine.
Many phosphorus chemical shifts have been reported relative to 850%H 3PO
4
as the standard.
APPENDIX
Solution of the
Secular Determinant for
the NMR Coupling of
the AB Spin System
To evaluate these matrix elements for a second-order, AB system in Section 7-20
we need to know how the IA 'B operator works. This problem is simplified by
defining the so-called raising and lowering operators.We shall have occasion to
use these operators in other problems. Remember that
2 2 + jY2 + Z2 (G_1)
but that we can only find simultaneously 12 and the component in one direction.
The raising and lowering operators, 1 and L_, are defined by taking linear
combinations of !X and ly such that:
Z, = Zx + iZy (G-2)
L = ix - ily (G-3)
These operators have the property that when they operate on a wave function
I, mi> (i.e., one defined by quantum numbers I and m) we get:
1, IM 1> = [I(I + 1) -mr(mi + 1)]1/2' I,m + 1>

L_ I,mI) = [(I(I + 1)- mr(m - 1)]i 2|I,m - 1>
Then we find that these operators work on I > and |#> as follows:
Zx>= 0 (we cannot raise a +/2 spin by 1 when I= 2 )
1 |#> -0 (a 2 spin cannot be lowered by 1 when 1 =/2)
The !A B operator for a two-spin AB system is given by:
IA - I- AIZB ±XA XB +IYA YB
740 = ZZA 'ZB ± (/2)(+A' -B + 1-A'+B) (G-4)

Appendix G Solution of the Secular Determinant for the NMR Coupling of the AB Spin System 741
Equation (8-18) in terms of the raising and lowering operators can be derived
by solving equations (8-16) and (8-17) for 7, and 1, and substituting this result
into the equation for !A iB in terms of X, Y, and Z components. In operating
on the basis set (o, etc.) with the 'A 1B operator, the lA spin operator acts only
on the A nucleus (the first spin function listed) and the 'B Operator acts only on
the B nucleus (the second one listed). Accordingly,
IZA P (/210
while
The general Hamiltonian for the coupled AB system, equation (8-15), becomes
equation (G-5) when expressed in terms of the raising and lowering operators.
HAB V0 (I - a)!ZA - 0( BZB + JAB ZA 'ZB
+ ( 2)JAB( +A -B + A +B) (G-5)
Now we return to the evaluation of the matrix elements in the secular

determinant given earlier. To evaluate (<S I H I oo>, we need to evaluate the effect
of ft on I cc> and then simply multiply by (z 1. Thus, we shall proceed by first
evaluating ft 22>, fI a#>, ft I fl2>, and h I##>.
HIc=> [-v,(l -A2) 0 - 6B) 2 +( 4 )JI ">
The ('/4 )J arises from IZA * Z1BI"> for both i+AI-B I mc> and AUB Cot> equal
zero, since I+A I a) = 0 and I+ B I > = 0. Rearranging the above expression for
Ha>,we have
HIa>= vo 1 +'7B A+ + |j>
Io#> = -ve(1 - 0A) 2 v(O1 -UB)(-

B ill,> J[0 + I3l>
= 1 1
Iv(1 = 212v4aA
- ') -V(1 - UB) 24- }/3c + J2 I[tfl> ±0]
=t33vV1 - 0A B)( IJ]

- VO(i C
A2 tB)±+ J]3>
With these quantities evaluated, it is a simple matter to multiply them by
22, afl, etc., and thus to evaluate all the matrix elements in the secular determinant.
742 Appendix G Solution of the Secular Determinant for the NMR Coupling of the AB Spin System
For example, (fl I ft I > ) = 0 + '/2 J. The results for all the matrix elements are
contained in Table G-1. The result of k Ia7> is seen to yield Ia1> back again,
so this wave function is an eigenfunction of the Hamiltonian, as is |##>. However,
I)cx> and fl> are not eigenfunctions. Accordingly, we find off-diagonal values
of ('/2 )J connecting af#and #. The determinant is block diagonalized, with two
frequencies given as:
SvoL-1+ I A + B] + (G-6)
E41 1 J
2 B] +4 (G-7)
h =vO 1 2A
The energies of E 2 and E3 are obtained by solving the 2 x 2 block. This solution
can be simplified by making a few definitions. The definitions to be made are
not obvious a priori choices. They come with experience in solving determinantal
equations and trying to relate the results to observed spectra. We will substitute:
TABLE G-1. Evaluation of the Matrix Elements for the Coupled AB System
xx v -1+ +A +a,00]
J E J
xp 0 + 2 vO((A - GB) 4 - 2
J J E
#2 f 0
02 - B - 4
2 vo(onA- 'TB) - h0 0
pp 0 0 0 v -42 2 -
)v(oA- ) C cos 20
and
J
- C sin 20
2
where C = ( 1/2)(j 2 + A2 )" 2. The geometrical relation of these defined quantities

is shown in Fig. G-1. With these definitions the 2 x 2 block of the secular
determinant becomes:
2C
1 E
-C cos 20 - J- C sin 20
/204 h
27 =0
A 1 E
C sin 20 C cos 20 - - J
4 h
Appendix G Solution of the Secular determinant for the NMR Coupling of the AB Spin System 743
The solutions to this determinant are
-2 = --- - C (G-8)
h 4
E3 1
= - J +C (G-9)
To get the wave functions for these two states, we substitute the energies one at
a time into the determinant given above. The matrix corresponding to this
determinant (this matrix has the same form as the determinant) is multiplied by
a column matrix (vector) of the coefficients. The product is zero. Expansion
produces the two resulting, linearly dependent equations, which can be solved
for C, and C 2 when the normalization requirement is imposed. In general, this
is done for each energy. The procedure is similar to that used in Chapter 3 when
Huckel theory was discussed. The wave functions 02 and 0, so obtained are
listed below along with 0, and 04.
1 la) (G- 10)
02 cos 6|l> - sin 0|#2> (G- 11)

0 3 sin 0|1> + cos 0f#> (G- 12)
04 = |##) (G- 13)
Index
A, symmetry, 169 Bending vibration, 92 local effects, 232-238

AB molecule spectrum, relative magni- Benzene, molecular orbital, 116 measurement, 229-232
tudes of J and delta, 263-264 Beta calculation, OhNi(II) complex, 438- remote effects, 238-240
AB spin system, nuclear magnetic reso- 443 Chromophore, heme, 171
nance coupling, solution of secular Bimetallic system, nuclear magnetic reso- Circular dichroism, 137-141
determinant, 740-743 nance, 543-552 Closed shell system
Absorption band, effects giving rise to, Bloch equation, 218-220 antishielding, 615
154-156 Block diagonalized, 34 shielding, 615-616
Activation enthalpy, nuclear magnetic res- Blue shift, 118-119 CO 2 molecule
onance, 290-295 Boltzmann constant, 95 symmetric stretching vibration,
Adiabatic electron affinity, M-, 666 Bond dissociation energy, force constant, polarizability change, 165
Allyl, intermediate neglect of differential 152 vibrations, 155
overlap, 80-81 Bond order, 73, 74 Cobalt-substituted carbonic anhydrase, 543
Allyl radical Bond strength, frequency shift, 194-202 Cobalt(II)
pi system, 69 Bonding parameter, 450-458 diastereoisomerism, 535-537
wave function, 76-78 tetragonal complex, 450-452 diastereotopism, 535-537
Alternately double bonded hydrocarbon. Born-Oppenheimer approximation, 94 Co(Meacacen), electron paramagnetic reso-
molecular orbital theory, 83-84 Bragg's law, 690 nance spectrum, 563
Ammonia, ultraviolet photoelectron spec- Bravais lattice, 703 Complete neglect of differential overlap
troscopy, 673 (CNDO), 80
3
Angular momentum, 213 C magnetic resonance, 319-323 Concentration, determination, 99-103
Angular overlap model, 452-458 C2, point group, character table, 114 Configuration interaction, 117-118
Anharmonic vibration, 149-150 Cambridge Crystallographic Database, Conjugate, 22
Anion, ambidentate, 198 689 Constant-acid, constant-base frequency
Anionization, secondary ion mass spec- Cambridge Structural Database, 710 shift-enthalpy relations, 195
trometry, 678 Carbon dioxide, fundamental vibration Contact shift, 507-508
Anisotropy modes, 155 factoring, 512-514
g value, 380-383 Carbonyl compound, 131 spin density, 514-518
hyperfine coupling, 383-390 Cartesian basis set, 36 Contrast agent, nuclear magnetic reso-
A.O. basis set, nitrite ion, 61 Cartesian coordinate system, 30-43 nance, 530-531
Appearance potential, 655, 664-665 Cationization, secondary ion mass spec- Coordinate, normal, 93
Asymmetric stretch, 154 trometry, 678 Coordinate system, NH 3 location, 63
Asymmetric top, 190-191 Character table, 35-36, 41-43 Coordination
Asymmetry parameter, 269, 606 C2 , point group, 114 donor molecule spectra change. 196-198
Atom, equivalent, 12-13 double group, 724-726 spectra and symmetry changes, 198-
Atom r, qr, electron density, 72-73 symmetry group, 712-723 202
Atomic force microscopy, 680-681 Charge-transfer spectrum, solvent polarity, Copper (II) complex, 590
Atomic transition, 92-95 135-137 Coupling, quantum mechanical descrip-
Auger electron spectroscopy, 679 Charge-transfer transition, 127-128 tion, 259-263
Chemical exchange, spectral line width, Coupling combination band, symmetry,
Band assignment, 156-160 95-98 176
criteria, 118-119 Chemical ionization, 654 Coupling constant, determining signs,
Base-iodine addition compound, molecular Chemical shift 268-269
orbital, 134 data conversion, 739 Crystal, 695-698
Base-iodine solution, spectra, 133 fluorine, 235 classes, 701
Basis set, 36 interpretation, 232-240, 241-243 growth, 695
745
746 index
Crystal (continued) Double resonance electron paramagnetic Ethylpyridine, mass spectra of 3 isomers,
liquid-diffusion vessel, 695 resonance, 594 658
vapor diffusion, 695 Double resonance experiment, 267-268 Eulerian cradle. 698-699
mounting, 697-698 Expectation value, 53, 504-507
selection, 696-697 e2Qq, nuclear quadrupole resonance spec- Extended X-ray absorption fine structure,
size, 696 troscopy, 616-617 681
X-ray absorption, 696-697 crystal lattice effects, 617
Crystal field, 416-432 data interpretation, 616-617 Far infrared vibrational frequency correla-
Cu 2 Co 2 superoxide dismutase, 545 Effective spin Hamiltonian, 363 tion chart, 188-189
Cu 2Co 2SOD derivative, 'H nmr spectrum, Eigenvalue, 53, 81-82 Fast atom bombardment, 654
546 Eigenvector, 81-82 Fast chemical reaction, spectrum effect,
Curie relaxation, equations, 523-524 Electric field gradient, 605 257-258
Curie temperature, 471 molecular structure, 612-616 Fast nuclear relaxation, two-dimensional
Cytochrome C peroxidase, X-band spec- Electric quadrupole, intensity, 127 spectrum, 532-533
7
trum, 585 Electromagnetic radiation, plane-polarized, 5 Fe
91 magnetic splitting, 634
d orbital, 126-127 Electron correlation, 80 non-cubic electronic environment, 632
d orbital energy, ligand, 416-419 Electron density, 72-73 quadrupole splitting, 634
symmetry, 419-427 atom r, q;, 72-73 Fe 2S2 ferrodoxin, 'H nmr spectrum, 546-
d orbital energy level, distortion, 443-447 Electron-electron interaction, 409-413 548
di, 578-579 term symbols, 409-410 oxidized, 547
d2, 579 Electron ionization, 654 reduced, 547
d3 , 579-581 Electron loss, secondary ion mass spec- Fe 3(CO)1 2
d4 , 581 trometry, 678 M6ssbauer spectroscopy, 640
d', tetragonal field, 585 Electron paramagnetic resonance (epr) structure, 640
d5 high spin, 583-586, 590 spectroscopy, 360-399 FeFe(CN)r, M6ssbauer spectroscopy, 629
d' low spin, 581-583 anisotropic effects, 380-399 Fermi contact contribution, 576
d9 , 590-591 applications, 397-399 Fermi contact coupling, 501
Decomposition formula, representation, Co(Meacacen), 563 Fermi resonance, 155-156
44-46 line width, 394-396 symmetry, 176
Degenerate representation, 38-40 metal cluster, 591-594 Ferrimyoglobin, X-band spectrum, 585
Delta, formal charge, 73 principles, 360-363 Field desorption, 654
Depolarization ratio, 169 spectrum presentation, 368-370 glutamic acid, 663
Depolarized line, 169 transition metal ion complexes, 559- Field gradient
Diamagnetic term, 234 594 pyridine, 619
Diamagnetism, 471-473 first-row survey, 578-591 semiempirical approach, 615
Diastereoisomerism, 535 triplet state, 390-392 Field gradient q, 269
cobalt(II), 535-537 Electron relaxation, spin-orbit coupling, Field gradient tensor, principal axis sys-
Diastereotopism, cobalt(II), 535-537 578 tem, 605-606
Diatomic molecule, 190 Electron spin, 503-507 Field ionization, 654
electronic energy level, 109-110 Electron spin state, zero-field splitting, glutamic acid, 663
energy state, 94 573 Field ionization technique, 662-663
Morse energy curve, 109-110 Electronic absorption spectroscopy, 109- Fine splitting, 572
vibrational energy level, 110 143 Fingerprinting, 106, 130-132, 184-186
Differential overlap, intermediate neglect, applications, 130-137 carbonyl compound, 131
78-83 nomenclature, 113-116 inorganic system, 132
Diffraction pattern, 693 Electronic energy level, diatomic mole- mass spectrometry, 657-659
Diffractometer, 697, 698 cule, 109-110 saturated molecule, 130-131
Dimeric complex, 'H nmr, 544-545 Electronic relaxation time, 527-530 First overtone, 151
Dimeric copper adenine complex, 592 Electronic spectrum Fluorine, chemical shift, 235
1,2-Dimethyl imidazole, Mossbauer spec- energy curve, I 1-112 Force constant, 149, 151-153
troscopy, 635 oxo-bridged dinuclear iron center, 458- bond dissociation energy, 152
Dipolar hyperfine coupling, electrons in d 459 Force constant matrix, compliance con-
orbitals, 575 structural evidence, 447-450 stants, 160
Dipolar relaxation, equations, 522-523 Electronic state, change, 92 Formal charge, 73, 75-78
Dipole moment, 13-14 Electronic transition, intensity, 120-130 delta, 73
Direct dipolar coupling, 340-341 Electrons in d orbitals, dipolar hyperfine Formaldehyde
Direct lattice, monoclinic cell, 690-692 coupling, 575 molecule shapes, 114
Direct product, 46 ENDOR experiment, 594 spectrum, 123-124
Distance, polarizability, 167 Energy curve, electronic spectrum, Ill- Fourier transform epr, 594
Distortion, d orbital energy level, 443-447 112 Fourier transform ion cyclotron resonance
Doppler energy, 627 Energy state, diatomic molecule, 94 technique, 665-667
Double group, 427-429 Equivalent atom, 12 Fourier transform nmr, 309-319
character table, 724-726 Et4 N', ethyl group disorder, 710 experiment optimization, 314-315
Double resonance, nuclear quadrupole res- Ethylmethyl sulfide, Mossbauer spectros- multipulse methods, 328-340
onance spectroscopy, 620-622 copy, 635 other nuclei, 323-325
index 747
principles, 309-314 Hybridization of ligand, 577 lIeducible representation, 34-35
relaxation, 326-328 Hydrocarbon system, pi orbital, 68 wave function, 59-60
selective excitation and suppression, Hydrogen atom, 363-368 Isomer shift, Mbssbauer spectroscopy,
333-334 Hydrogen-bonded phenol, infrared spectra, 630-631
sensitivity-enhancement methods, 331- 194 Isomolar solution, Job's method, 106
332 Hydrogen bonding solvent, blue shift, Isosbestic point, 103-106
spectral density, 326-328 119 Isotope, mass spectrum, 660-662
T1 measurement, 315-316 Hyperfine coupling, 571-576 Isotropic system, hyperfine coupling con-
Franck-Condon factor, 672 anisotropy, 383-390 stant, 374-379
Franck-Condon principle, 11I Hyperfine coupling constant, isotropic sys-
Free induction decay, 621 tem, 374-379 Jahn-Teller effect, 429-430
Free ion, spin-orbit coupling, 413-416 Hyperfine splitting, nitroxide, 399 Job's method, isomolar solution, 106
Free ion electronic state, 409-416 Hypsochromic shift, 118-119
Frequency shift, bond strength, 194-202 Kramers' doublet, 559
Fundamental frequency, 151 /, properties, 224-225 rhombic symmetry, 586
Fundamental vibration frequency, 150 Identity, 2 Kramers' rule, 559
Identity element, 19
g -> u, 123 Infrared group frequency range, 186-187 Lanthanides, 510-512
g value, anisotropy, 380-383 Infrared line, symmetry, 172-176 Laser desorption, 654
Gamma ray, 626, 627 Infrared spectra, 194 Lattice plane
Gas, spectra, 186-192 Infrared spectroscopy Miller indices, 690-691
Geometric transformation, 30-34 inorganic structure, 192-194 direct, 691
Glide plane, 11 procedures, 179-184 reflection, 691
Glutamic acid, 663 Inorganic anion, characteristic absorption Lattice vibration, 183
Goniometer, 698 maxima, 132 Ligand, d orbital energy, 416-419
Goniometer head, 697 Inorganic Crystal Structure Data Base, 710 symmetry, 419-427
Gross atomic population, 75 Inorganic material, infrared group frequen- Ligand hybridization, 577
Group H order, 20 cy range, 186-187 Ligand hyperfine coupling, 576-578
Group multiplication table, 19-25 Inorganic structure, 192-194 Line width
element classes, 23-25 Inorganic system, fingerprinting, 132 chemical exchange processes, 561
properties, 19-22 Integral zero, 121 electron paramagnetic resonance spec-
Group theory, 18-35 Integrated intensity, 120 troscopy, 394-396
element rules, 19 Intensity Linear combination of atomic orbitals-
generation of symmetry combinations, electric quadrupole, 127 molecular orbital method, 54
18 magnetic dipole, 127 Linear molecule
zero quantum mechanical integrals, 18 spin-orbit, 124-126 bending mode, 154
Group vibration, 160-162 spin-orbit coupling, 124-126 normal modes of vibration, 153
limitations, 161 vibronic coupling, 124-126 rotational mode, 154
Interaction force constant, 157 Linear polyatomic molecule, 190
'H nmr spectrum Interatomic ring current, 241 Local symmetry, 124
Cu 2Co 2SOD derivative, 546 Intermediate neglect of differential over- Longitudinal relaxation, 216
Fe 2S2 ferrodoxin, 547 lap, allyl, 80-81 Low energy electron diffraction, 678-679
H2, equivalent orientations, I Intervalence electron transfer band, 459-
H20 molecule, internal coordinates, 157 461 M-, adiabatic electron affinity, 666
Hahn spin-echo experiment, 328 Inversion center, 2 Magnetic behavior, types, 471-476
Harmonic oscillator approximation, 156 Iodine Magnetic coupling, metal ion cluster,
Harmonic vibration, 149-150 molecular addition compound, 133- 430-432
Hockel procedure, 68-72 135 Magnetic dipole, intensity, 127
extended, 74-78 spectra, 133 Magnetic field
molecular orbital, 68-72 Ion pair, N-methylpyridinium iodide, 136 magnet in, 213-214
Heme, chromophore Ion pairing, 538 nuclear quadrupole resonance spectros-
ultraviolet (Soret), 171 Ionization. 650-682 copy, 611-612
visible, 171 Ionization efficiency curve, 664 T state, 570
Heme Raman spectrum, 170-171 Ionization potential, 664-665 Magnetic splitting, 57 Fe, 634
Hemin-imidazole-cyanide, 538-539 Iron(II) complex Magnetic susceptibility, measurement,
Hemoglobin, schematic formula, 582 Massbauer spectroscopy, 638 490-491
High melting solid, vapor over, 663 partial quadrupole splitting parameter. Magnetism, 469-494
High resolution electron energy loss spec- 644 high spin-low spin equilibrium, 489-
troscopy, 679-680 Iron(III) complex, Mossbauer spectros- 490
High resolution infrared spectrum, copy, 638 intramolecular effects, 486-489
Mo(CO)5Br, 201 Iron (III) complex susceptibility measurement applications,
High resolution nmr, solids, 347-348 spectrum, 584 483-486
High spin-low spin equilibrium, magne- splitting of energy levels, 584 Magnetization vector, 215-216
tism, 489-490 Iron porphyrin, spin delocalization. 539- Magnetocircular dichroism, 141-143
Higher-state mixing. 117 543 Marginal oscillator, peak, 609
Hooke's law constant, 149 Iron-sulfur compound, center shift. 641 Mass spectrometer, 653
748 Index
Mass spectrometric shift rule, 660 Mossbauer spectroscopy, 626-645 solid studies, 341-342
Mass spectrometry, 650-665 applications, 635-645 thermodynamic data evaluation, 290-
fingerprint, 657-659 1,2-dimethyl imidazole, 635 291
instrument operation, 650-665 ethylmethyl sulfide, 635 NMR coupling, AB spin system, solution
interpretation, 659-660 Fe 3(CO) 12 , 640 of secular determinant, 740-743
isotope, 660-662 FeFe(CN) 6 , 629 NMR kinetics, 201-309
molecule high energy electron combina- iron(II) complex, 638 applications, 297-300
tion, 655-657 iron(III) complex, 638 NMR spectroscopy, 211-270
presentation, 650-665 isomer shift, 630-631 quantum mechanics, 224-227
Matrix, 27-30 I-methyl imidazole, 635 second-order systems, 265-266
Matrix formulation, molecular orbital non-cubic electronic environment, 632 slow passage experiment, 220-224
calculation, 56-57 piperidine, 635 transition, 217-218
Matrix multiplication, 27-28 principles, 626-629 Nuclear Overhauser effect, 306-309, 531-
cis MB 4A2 pyridine, 635 532
coordinates, 643 quadrupole interaction, 631-633 Nuclear quadrupole interaction, 392-394
geometries, 643 Nuclear quadrupole moment, 605
trans MB 4A2 3N-6(5) rule, 153-154 Nuclear quadrupole resonance (nqr) spec-
4
coordinates, 643 1 N quadrupole transition, pyridine, 619 troscopy, 604-622
geometries, 643 Naphthalene, 707 applications, 616-620
Metal ion cluster atomic coordinates, 708 double resonance, 620-622
electron paramagnetic resonance spec- Neighbor anisotropy, 238-239 e2Qq, 616-617
trum, 591-594 Neel temperature, 471 crystal lattice effects, 617
magnetic coupling, 430-432 Net atomic population, 74 data interpretation, 616-617
1-Methyl imidazole, M6ssbauer spectros- Neutron diffraction, 701 magnetic field, 611-612
copy, 635 NH 3 location, coordinate system, 63 structural information, 618-620
N-Methylpyridinium iodide, ion pair, 136 Nickel, planar-tetrahedral equilibrium, Nuclear relaxation, paramagnetic system,
Microstate configuration, 409 534-535 519-524
Microwave spectroscopy, 177-179 Nitrite ion Nuclear spin quantum number, 626
Miller indices, lattice plane, 690-691 A.O. basis set, 61
direct, 691 pi-molecular orbital, 63 O-S-O bending mode, 154
Mirror plane, 4-5 Nitrogen, ultraviolet photoelectron spec- Octahedral complex, vibration, 126
Mo(CO)sBr, high resolution infrared spectroscopy, 672 Octahedral iron (III) complex
trum, 201 Nitroxide, hyperfine splitting, 399 spectrum, 584
Molar absorptivity, 118-119 NO, photoelectron spectrum, 676 splitting of energy levels, 584
Molecular addition compound, iodine, NO 3, vibrational mode symmetry Octahedral symmetry, deviations, 582
133-135 C 2 v structure, 198 O field
Molecular orbital correlation chart, 198 electronic spectra survey, 433-438
base-iodine addition compound, 134 C, structure, 198 Tanabe and Sugano diagram, 732-734
benzene, 116 Non-diagonal representation, 36-40 OhNi(II) complex
calculations, 68-72 Non-equivalent proton, 246 beta calculation, 438-443, 735-737
Huckel procedure, 68-72 Non-linear polyatomic molecule, 190-192 D. calculation, 438-443, 735-737
projecting, 60-68 Nonlinear molecule One electron molecular orbital, 54
Molecular orbital calculation, matrix bending mode, 154 Operator, 52-56
formulation, 56-57 normal modes of vibration, 153 OPF3, vibration, fundamental frequencies,
Molecular orbital theory, 52-85 rotational mode, 154 193
alternately double bonded hydrocarbon, Normal coordinate, 93 Optical activity, 13
83-84 Normal coordinate analysis, 156-160 Optical rotary dispersion, 137-141
Molecular structure Normal mode, 92 Orbital, nitrite ion, 63
electric field gradient, 612-616 Normal vibration, 92 Order of group h, 20
shift reagent determination, 550 NSF 3 , vibration, fundamental frequencies, Organic material, infrared group frequency
Molecular transition, 92-95 193 range, 186-187
Molecular vibration Nuclear hyperfine splitting, 363-379 Orgel diagram, 427
radiation absorption, 150-151 more than one nucleus, 370-374 Oscillator strength, 120
symmetry, 172-179 Nuclear magnetic resonance (nmr) Overhauser effect, 306-309
Molecular weight determination, 662- activation enthalpy, 290-295 Overlap population, 74-75
663 bimetallic system, 543-552 Oxo-bridged dinuclear iron center,
Molecule contrast agent, 530-531 electronic spectrum, 458-459
asymmetric, 13 fluxional behavior, 300-305 Oxygen
dissymmetric, 13 intramolecular rearrangements, 300-305 ultraviolet photoelectron spectroscopy,
structures of excited states, 137 liquid crystal solvent studies, 342-347 674
Monoclinic cell lattice, 690-692 paramagnetic substances in solution, vertical ionization data, 675
Morse energy curve, diatomic molecule, 500-552 Oxyhemerythrin, resonance Raman analy-
109-110 quadrupolar nucleus, 319 sis, 171
Morse potential, 109 rate constant, 290-295
M6ssbauer emission spectroscopy, 635 reaction order determination, 295-297 P-branch, 191
M6ssbauer isotope, 637 relaxation, 227-243 p orbital, 126-127
Index 749
PA, bond order, 73-74 R-branch, 191 cationization, 678

Parallel band, 190 Radiation electron loss, 678
Paramagnetic compound, properties, 500- allowed transitions, 95 sputtering, 678
503 corresponding energies, 91-92 vs. mass number/ion charge, 677
Paramagnetic Mbssbauer spectrum, 633- forbidden transitions, 95 Self-consistent field, 79
635 nature, 90-91 Shift reagent, 549-552
Paramagnetic substances in solution, nmr, selection rules, 95 Simple harmonic motion, 149
500-552 wave numbers, 92 Single crystal study, polarized, 129-
Paramagnetic system, nuclear relaxation, Radiation absorption, molecular vibration, 130
519-524 150-151 Site symmetry, 183
Paramagnetic term, 233 Raman line SO 2 , infrared spectrum, 154
Paramagnetism, simple system, 473-476 depolarized, 168-169 Solvent, Z-values, 136
Partial quadrupole splitting parameter, polarized, 168-169 Solvent polarity, charge-transfer spectrum,
iron(II), 644 symmetry, 172-176 135-137
Peak assignment, T,, 317-319 Raman spectroscopy, 162-171 Space symmetry, 11
Perturbation theory, 57-59 inorganic structure, 192-194 Spectral density, Fourier transform nmr,
Phenol, infrared spectra, 194 procedures, 184 326-328
Phonon mode, 183 selection rules, 164-168 Spectral line width
Photoacoustic Fourier transform spectros- Rapid nuclear quadrupole relaxation, 269 chemical exchange, 95-98
copy, 182 Rate constant, nmr, 290-295 relaxation, 95-98
Photoelectron spectroscopy, 667 Reciprocal lattice, monoclinic cell, 690- Spectroscopy, 90-106
Photoemission, 669 692 applications, 98-106
Pi orbital, hydrocarbon system, 68 Reciprocal space, 691-692 See also Electronic absorption spectros-
Pi system, allyl radical, 69 Reflectance technique, 181 copy
Piperidine, M6ssbauer spectroscopy, 635 Reflection, 691-692 Spherical top, 190-191
Planar-tetrahedral equilibrium, nickel, lattice plane, 691 Spin delocalization, iron porphyrin, 539-
534-535 Relaxation, 215-216 543
Planck's constant, 150 Fourier transform nmr, 326-328 Spin density, contact shift, 514-518
Plasma desorption, 654 nmr, 227-243 Spin echo, 329-331
Point group, 8-11 spectral line width, 95-98 Spin Hamiltonian, 396-397, 573-574
symmetry element, 9 Relaxometry, 524-527 Spin-lattice relaxation, 216
Polarizability Representation, 30-34, 43-44 Spin-orbit coupling, 117, 483-486
distance, 167 decomposition formula, 44-46 electron relaxation, 578
vibration, 165-167 degenerate, 38-40 free ion, 413-416
Polarized absorption spectra, 128-130 irreducible, 34-35 intensity, 124-126
Polarized line, 169 non-diagonal, 36-40 large, 569-571
Polyatomic molecule, vibration, 153-162 Resonance Raman analysis, T state, 571
Product, 19 oxyhemerythrin, 171 Spin-polarization, 375
Product ground state wave function, 84- Resonance Raman spectroscopy, 170-171 results, 576
85 Rhombic symmetry, Kramers' doublet, Spin saturation labeling, 305-306
Projection operator, 60 586 Spin-spin coupling
Protein Data Bank, 710 Rotating axis system, 214-215 bond nature, 247-249
Protonation, 119 Rotating frame, 214-215 bond number, 247-249
Pseudocontact shift, 508-510 Rotation axis, 2-4 scalar mechanisms, 249-251
factoring, 512-514 dissymmetric, 13 structure determination, 252-257
Pyrene, 707 improper, 5 Spin-spin relaxation, 216
Pyridine proper, 2 Spin-spin splitting, 243-270
field gradient, 619 Rotation-reflection axis, 5-7 spectrum effect, 243-246
Mbssbauer spectroscopy, 635 Rotational line, Zeeman splitting, 179 Spin temperature, 621
"N quadrupole transition, 619 Rotational Raman spectrum, 179 Spin-tickling experiment, 267-268
Rotational state, 94 Sputtering, secondary ion mass spec-
Q-branch, 190, 191 trometry, 678
Quadrupolar nucleus, nmr, 319 S=1/2 system, orbitally non-degenerate Square matrix, 27
Quadrupole energy level ground states, 565-569 Square planar Schiff base ligand, 590
axially symmetric field, 607 Saturated molecule, fingerprinting, 130- Sternheimer effect, 615
spherical field, 607 131 Stokes line, 169
Quadrupole interaction, Mbssbauer spec- Scanning tunneling microscopy, 680-681 Stretching vibration, 92
troscopy, 631-633 Scattering factor curve, 694 Structure determination, spin-spin cou-
Quadrupole moment nucleus, 269-270 SCF-INDO, 78-83 pling, 252-257
57
Quadrupole splitting, Fe, 634 Screw axis, 11, 703 Sublimation heat, 663
Quadrupole transition, energies, 607- Second-order spectra, 259 Sulfur dioxide, three fundamental vibra-
611 Second overtone, 151 tions, 155
Quantum mechanics Secondary ion mass spectrometry, 654, Superparamagnetism, 491-494
nmr spectroscopy, 224-227 677-678 Superregenerative oscillator, peak, 608-
symmetry, 59-68 aluminum target, 677 609
Quasi-equilibrium theory, 660 anionization, 678 Surface science technique, 667-681
750 index
Susceptibility measurement applications, Through-space dipolar coupling, 501 Vibrational energy state, 92
magnetism, 483-486 Total ionization unit, 652 Vibrational mode symmetry, NO 3
Symmetric stretch, 154 Transition metal ion, 409-462 C 2 , structure, 198
Symmetric stretching vibration, CO 2 g-values, 564-570 correlation chart, 198
molecule, polarizability change, Transition metal ion complexes, epr spec- C, structure, 198
165 trum, 559-594 Vibronic coupling, intensity, 124-126
Symmetric top, 190-191 first-row survey, 578-591
Symmetry Transition moment integral, 120-122 Wave function, 52
center, 2 Transition probabilities, 225-227 allyl radical, 76-78
coupling combination band, 176 Transmission infrared, 181 irreducible representation, 59-60
definition, 1 Transverse relaxation time, 216 mixing p-character, 127
Fermi resonance, 176 Triplet state, epr spectroscopy, 390- properties derived from, 72-74
infrared line, 172-176 392 symmetrized, 84-85
plane of, 4-5 Two-dimensional nmr, 334-340
quantum mechanics, 59-68 Two-dimensional spectrum, fast nuclear X-ray
Raman line, 172-176 relaxation, 532-533 atom and structure scattering, 693-695
X-ray crystallography, 701-710 diffraction, 690
molecular vs. crystallographic, 707 u -- g, 123 X-ray absorption near-edge structure,
See also Space symmetry Ultraviolet photoelectron spectroscopy, 681
Symmetry element, 2-7 671-677 X-ray crystallography, 689-710
equivalent, 12-13 ammonia, 673 area detectors, 700-701
point group, 9 nitrogen, 672 avoiding crystallographic mistakes, 705-
Symmetry group, character table, 712-723 oxygen, 674 707
Symmetry operation, 12 Ultraviolet radiation, 94 computers, 700
Symmetry species of state, 115 Unit matrix, 27 crystallographic data, 710
Synchrotron radiation, 701 Unpaired electron density, 375 diffraction equipment, 698
diffractometer data collection, 698-
T state Van Vleck's equation, 476-483 700
magnetic field, 570 Variation energy, 55 future developments, 700-701
spin-orbit coupling, 571 Variational principle, 55 methodology, 698-700
TI, peak assignment, 317-319 Vector, 25-26 principles, 690-695
Tanabe and Sugano diagram, 426 Vibration, 92-94 quality assessment, 707-709
O field, 732-734 bending, 92 reflection conditions, 704
TdCo" complex, calculation of delta(IODq) harmonic, 149-150 space-group determination, 704-705
and beta, 735-737 nomenclature, 172 space groups, 702-703
Tetragonal complex, bonding parameter, normal, 92 symmetry, 701-710
450-452 normal modes for common structures, molecular vs. crystallographic, 707
pi, 450-452 727-731 symmetry elements, 704
sigma, 450-452 NSF 3, fundamental frequencies, 193 vs. neutron diffraction, 701
5
Tetragonal field, d octahedral complex, 126 X-ray photoelectron spectroscopy, 668-
strong, 585 OPF 3, fundamental frequencies, 193 671
weak, 585 polarizability, 165-167
Thermal desorption, 654 polyatomic molecule, 153-162 Z-axis compression, in D4h, 564
Thermodynamic data evaluation, nmr, stretching, 92 Zeeman Hamiltonian, 565
290-291 symmetric, 165 Zeeman splitting, rotational line, 179
Thiocarbonyl compound See also Specific type Zero-field splitting, 391, 571-576
carbonyl compound, 131 Vibrational energy level, diatomic mole- anisotropic, 573
spectra, 131 cule, 110 electron spin state, 573
MIT LIBRARIES
3 9080 02682 3754
SELECTED CONSTANTS*
Value
Symbol Quantity cgs Units SI Units
k Boltzmann constant 1.3807 x 101 erg K' 1.3807 x 10-23 J K

h Planck constant 6.6262 x 10-27 erg s 6.6262 x 10" J s
c speed of light in vacuum 2.9979 x 1010 cm s-' 2.9979 x 10' m s-
N Avogadro constant 6.0220 x 1023 Mol-1 6.0220 x 1023 mol
1.6022 x 10-20 emu
e elementary charge or 1.6022 x 109 C (sA)
4.8032 x 10"0 esu
amu atomic mass unit 1.6606 x 10-24 g 1.6606 x 10-27 kg
f(pB) Bohr magneton 9.2741 x 10-21 erg G-1 9.2741 X 10-24 J T-1
#, Bohr magnetin 4.6685 x 10-1 cm-1 G-1
#N nuclear magneton 5.0508 x 10-24 ergG '5.0508 x 10-27 J T
m, electron rest mass 9.1095 x 1021 g 9.1095 x 10 3 kg
mP proton rest mass 1.6726 x 10-24 g 1.6726 x 10-27 kg
pe electron magnetic moment 9,2848 x 10-21 erg G-1 9.2848 X 10-24 J T
YH proton magnetogmic ratio 2.6752 x 103 kg-1 sA
3.14159 26535 89793 23846 26433 83279 50288
Tstands forTsla(kgs-A-');CostandsrforCoulomb;S = second;A ampere; 10-9erg
1 gC sA)
CONVERSION FACTORS
1cal = 4.184 J (Joules)
1 Hz = 6.6262 x 10-" J
1 K = 1.3807 x 10- 27 j
1 kWh = 3.6 x 106 J
Energy Conversion Factors
erg/molecule ev cm 1 cal/mole
erg/molecule...... 1 6.242 x 1011 5.036 x 10" 1.439 x 1016

ev..................... 1.602 x 10-12 1 8,067 23,060
cm-1 ................ 1.986 x 10-16 1.2396 x 10~4 1 2.858
cal/mole............ 6.949 x 10-17 4.338 x 10-' 0.3499 1
PROPERTIES OF SELECTED NUCLEI
Relative Sensitivity Magnetic
NMR for Equal Numbers of Moment, pNia, Electric
Frequency in Nuclei in multiples Quadrupole Anisotropic Isotropic
MHz at a Natural at at of nuclear Spin, I, Moment, 0 Hyperfine Hyperfine
7.0463 T Abundance constant constant magneton in multiples in multiples Coupling, B Coupling, A.
24 2
Isotope Field field frequency (en 47nmc) of h/2n of 10- cm in MHZb in MHz,
300.000 99.9844 1.000 1.000 2.79270 1.420

.2H
46.051 1.56 x 10-2 9.64 x 10 - 0.409 0.85738 2.77 x 10-3 218
.6 Li 44.146 7.43 8.51 x 10' 0.392 0.82191 4.6 x 10-4 152*
7
- Li 116.590 92.57 0.294 1.94 3.2560 -4.2 x 10 2 402*
. 9 Be 42.160 100 1.39 x 10-2 0.703 1.1774 2 x 10-2 672
- 10B 32.239 18.83 1.99 x 10-2 1.72 1.8006 0.111 17.8 -672
-"1B 96.251 81.17 0.165 1.60 2.6880 3.55 x 10-2 53.1 2,020
-"3C 75.432 1.108 1.59 x 10-2 0.251 0.70216 90.8 3,110
14
- N 21.611 99.635 1.01 x 10-3 0.193 0.40357 2 x 10-2 47.8. 1,540
-15N 30.398 0.365 1.04 x 10- 0.101 -0.28304 -67.1 -2,160
.170 40.670 3.7 x 10-2 2.91 x 10-2 1.58 -1.8930 -4x 10-3 -144 -4.628
.19F 282.231 100 0.834 0.941 2.6273 1515 47,910
-23 Na 79.353 100 9.27 x 10-2 1.32 2.2161 0.1 886*
.2sMg 18.358 10.05 2.68 x 10-2 0.714 -0.85471
-27 Al 78.172 100 0.207 3.04 3.6385 0.149 59 2,746
.29Si 59.595 4.70 7.85 x 10-2 0.199 -0.55477 -86.6 -3.381
-31 P 121.442 100 6.64 x 10-2 0.405 1.1305 287 10,178
-33
35
S 23.009 0.74 2.26 x 10 - 0.384 0.64274 -6.4 x 10-2 78 2,715
C1 29.395 75.4 4.71 x 10-3 0.490 0.82089 -797 x 10 2 137 4,664
37
- C1 29.395 24.6 2.72 x 10-3 0.408 0.68329 -6.21 x 10-2 117 3.880
- 39 K 24.467 93.08 5.08 x 10- 0.233 0.39094 231*
.43Ca 20.184 0.13 6.39 x 10-2 1.41 -1.3153
.4sSc 72.882 100 0.301 5.10 4.7491 1,833
-4 7Ti 16.910 7.75 2.10 x 10-3 0.659 -0.78712 -492
.49Ti 16.914 5.51 3.76 x 10-3 1.19 -1.1023 -492
-sV 78.864 -100 0.383 5.53 5.1392 0.3 2,613
. 53Cr 16.956 9.54 1.0 x 10- 0.29 -0.4735 -630
.s Mn 73.993 100 0.178 2.89 3.4610 0.5 3,063
57
Fe 9.693 2.245 - <;0.05 450
. 59 Co 70.842 100 0.281 4.83 4.6388 0.5 3,666
60 Cot 4.6 5 x 10-2 4.3 3.0
611Ni 26.808 1.25 3.52 x 10-3 <0.25 1,512
. 6 3 Cu 79.515 69.09 9.38 x 10-2 1.33 2.2206 -0.15 4,952
.6 5Cu 85.183 30.91 0.116 1.42 2.3790 -0.14 5,305
.67 Zn 18.762 4.12 2.86 x 10-3 0.730 0.8735 1,251
.69 Ga 72.009 60.2 6.93 x 10-2 1.201 2.0108 0.2318
"7Ga 91.485 39.8 0.142 1.525 2.5549 0.1461
- 73 Ge 10.465 7.61 1.40 x 10-3 1.15 -0.8768 -0.2
75 As 51.380 100 2.51 x 10-2 0.856 1.4349 0.3 9,582
77 Se 57.203 7.50 6.97 x 10-3 0.191 0.5333 13,468
. 79 Br 75.160 50.57 7.86 x 10-2 1.26 2.0990 0.33 21,738
.81Br 81.018 49.43 9.84 x 10-2 2.2626 0.28 23,432 3
.85sRb 28.965 72.8 1.05 x 10-2 1.3483 0.31 1,012*
* Magnetic moment measured by nmr.

t Radioactive.
- Experimental hyperfine couplings as measured using the atomic-beam technique [See P. Kusch and H. Taub. Phys Rev., 75, 1477 (1949).]
a The magnetogyric ratio can be obtained from yN = pNI.
b Anisotropic hyperfine coupling given by

2
B =- h-'gNfNgfl(r 3)
5
where (r -3) is computed for a valence p electron from self-consistent-field wave functions. The couplings are such that the principal values of the trace,
tensor are respectively - 1, -1, and +2 times the number quoted.
Isotropic hyperfine coupling given by
8nr 2
A0 = - h- gNNgflk o)1
3
where W,0) is the value of the valence-shell, self-consisten tw e function at the nucleus of the neutral atom.
.11.11R. es-seA
S9
Magnetic
NMR Relative Sensitivity Moment, pNa Electric
Frequency in for Equal Numbers of in multiples Quadrupole Anisotropic Isotropic
MHz at a Natural Nuclei at of nuclear Spin, I, Moment, 0 Hyperfine Hyperfine
7.0463 T Abundance constant magneton in multiples in multiples Coupling, B Coupling, A0 ,
4 2
Itope Field % field (en/4nr mc) of h/2m of 10- cm in MHZb in MHz'
87 0.177 2.7415 3/2 0.15 3,417*

Rb 98.163 27.2
87Sr 13.001 7.02 2.69 x 10-3 -1.0893 9/2
89Y 14.697 100 1.17 x 104 -0.1368 1/2
91 9.4 x 10-3 -1.3 5/2
Zr 27.991 11.23
93
. Nb 73.328 100 0.482 6.1435 9/2 -0.4 + 0.3
95
. Mo 19.554 15.78 3.22 x 10-3 -0.9099 5/2 - - -3,528
. 9 7Mo 19.557 9.60 3.42 x 10-3 -0.9290 5/2 - - -3,601
9 - 6/2 - -
.9 Ru 10.169 12.81
101Ru 14.824 16.98 - 5/2 - -
. 1 0 3 Rh 9.442 100 3.12 x 10-5 -0.0879 1/2 -
i'Pd 13.728 22.23 7.79 x 10-4 -0.57 5/2 -
07 6.69 x 10~ -0.1130 1/2 - -3,520
.1 Ag 12.139 51.35
.1 09Ag 13.956 48.65 1.01 x 104 -0.1299 1/2 - - -4,044
- 1 "tCd 63.616 12.86 9.54 x 10-3 -0.5922 1/2
-1 3 Cd 66.548 12.34 1.09 x 10-2 -0.6195 1/2 - -
1 15
In 65.742 95.84 0.348 5.5072 9/2 1.161
"17 Sn 106.875 7.67 4.53 x 10-2 -0.9949 1/2
119 5.18 x 10-2 -1.0409 1/2 - -
Sn 111.817 8.68
"21 Sb 71.791 57.25 0.160 3.3417 5/2 -0.8
23
Sb 38.878 42.75 4.57 x 10-2 2.5334 7/2 -1.0 -
12 5
Te 94.790 7.03 3.16 x 10-2 -0.8824 1/2 - -
1271 60.021 100 9.35 x 10 -2 2.7939 5/2 -0.75 -
129Xe 80.981 26.44 2.12 x 10-2 1/2 -
17 7 Hf 9.361 18.39 6.38 x 10~ 7/2
79
1 Hf 5.609 13.78 2.16 x 10-4 - 9/2 - -
181Ta 35.910 100 3.60 x 10 -3 2.1 7/2 6.5 -
183w 12.48 14.28 7.20 x 10- 4 0.115 1/2 -
185 Re 67.541 37.07 0.133 3.1437 5/2 2.8 -
187 Re 68.233 62.93 0.137 3.1760 5/2 2.6
89
1 0s 23.276 16.1 2.24 x 10 -3 0.6507 3/2 2.0 - -
191Ir 5.156 38.5 3.5 x 10-5 0.16 3/2 ~1.2 -
93
1 Ir 5.614 61.5 4.2 x 10-5 0.17 3/2 ~1.0
95
1 Pt 64.447 33.7 9.94 x 10-3 0.6004 12
97
1 Au 5.138 100 2.14 x 10-5 0.136 3/2 0.56 -
199
Hg 53.481 16.86 5.72 x 10-3 0.4993 1/2 - - -
01 1.90 x 10-3 -0.607 3/2 0.5 -
2103 Hg
19.799 13.24
Ti 171.448 29.52 0.187 1.5960 1/2 -
20sTi 173.124 70.48 0.192 1.6114 1/2
207
Pb 62.765 21.11 9.13 x 10-3 0.5837 1/2 -
20 9
Bi 48.208 100 0.137 4.0389 9/2 -0.4
Electron 197.246 -2.85 x 108 1836 12-
2.00

Drago - Physicalmethodsforchemists 2e PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Drago - Physicalmethodsforchemists 2e PDF

Uploaded by

Copyright:

Available Formats

Physical

P.O. Box 13413

Copyright @ 1992, 1977

Text Typeface: Times Roman

Library of Congress Cataloging-in-Publication Data

OTHER SURFSIDE TITLES:

SOLUTION MANUAL - PHYSICAL METHODS FOR CHEMISTS by Ngai M.

APPLICATIONS OF ELECTROSTATIC-COVALENT MODELS IN

The revision of my earlier textbook, "Physical Methods in Chemistry," was

5-2 Relationship of Potential Energy Curves to Electronic Spectra 111

6-6 Normal Coordinate Analyses and Band Assignments 156

Nuclear Magnetic Resonance Spectroscopy-Elementary Aspects 211

The Quantum Mechanical Description of the NMR Experiment 224

Fourier Transform NMR 309

9-9 Nuclear Quadrupole Interaction 392

11-3 Van Vleck's Equation 476

Surface Science Techniques 667

17-1 Introduction 689

References Cited 711

Character Tables for Chemically Important Symmetry Groups 713

Character Tables for Some Double Groups 724

Calculation of A (1ODq) and # for Oh Ni" and Td Co" Complexes 735

Conversion of Chemical Shift Data 739

1-1 DEFINITION OF SYMMETRY Introduction

1-2 SYMMETRY ELEMENTS

The Center of Symmetry, or Inversion Center

The Rotation Axis

For many purposes, it is convenient to locate a molecule and its symmetry

The Mirror Plane or Plane of Symmetry

The Rotation-Reflection Axis; Improper Rotations C2

FIGURE 1-11 Improper rotation axis in the staggered form of ethane.

A particular orientation of the molecule methane is illustrated in Fig. 1-13.

FIGURE 1-13 Effect of the

solid symbols below.)

Next we shall consider some differences in improper rotation axes of even

S6 cannot be written any other way.

Se6 cannot be written any other way.

The complete set would normally be written:

S" = C.' and then a

The boldface indicates S5 operations that cannot be written as a single

TABLE 1-1. A Summary of Symmetry Elements and Symmetry Operations

identity E all molecules

reflection plane a H 20, BF3 (planar)

aFerrocene is staggered and possesses an S10 improper rotation axis.

1-3 POINT GROUPS

FIGURE 1-14 The C2-axes and two Ud planes in allene.

Point Group Symmetry Elementsa Examples

C2 one C2 axis H 202

C2 one C 2 axis and two a. planes H2O, SO2Cl2, SiCl2Br2

C3 one C3 axis and three a, planes NH3, CH 3 Cl, POCI,

D2 h three C2 axes all I, two u, planes, one a, plane, N 20 4 (planar)

D3h one C3 , three C2 axes I to C3, three a, BCl 3

D 2d three C2 axes, two ad planes, and one S4 H2C=C=CH2

Td three C2 axes I to each other, four C3 , six a, CH 4, SiCI4

'All point groups possess the identity element, E.

H 0-H A more exhaustive compilation of the important point groups is contained

1. Determine whether or not the molecule belongs to one of the special

The following flow chart provides a systematic way to approach the

S,, or S2 .and i only,

1-4 SPACE SYMMETRY

In describing the symmetry of an arrangement of atoms in a crystal (essential to

1-5 SOME DEFINITIONS AND APPLICATIONS OF SYMMETRY

Products or Combinations of Symmetry Operations